“A few days ago, everyone read about this year’s WalterCon being canceled, which left people with non-refundable, non-transferable [airplane] tickets (you can read Mikesum32’s reaction here) in their hands. Fortunately for them, an alternative has now been set up, and they will be able to still meet, in San Francisco, on August 11th. The venue? Picnix 16, a Linux gathering. The name? FalterCon 2007. Read on for my thoughts on this.”
The situation is unacceptable. Some conferences cost me ~$3000 USD to attend (most of that money spent on plane tickets and hotel). I would have been really pissed off if they got cancelled.
Things like that should never happen. I think the people behind Haiku should have made the conference anyway so the people who paid would have got something.
Bad day for Haiku.
Score: 17.
WTF???
Maybe setting up events, selling non-refundable tickets and then just canceling would make for a lucrative business.
How is it even remotely legal or ethical to sell non-refundable tickets?
I think the non-refundable tickets part refers to things like plane-tickets and other travel stuff which, under certain conditions, aren’t refundable.
I was referring to plane tickets.
As I noted elsewhere (Haiku’s official site, begroovy) I was this close to deciding whether or not I would be able to manage attending WalterCon this year with my current employment situation and othr things, with such short notice (less than a month, total) when it must be kept in mind that plane tickets, rental cars, rooms, etc. tend to become quite a bit more expensive if you try to reserve them 2 weeks or less before an event, unless the destination is so odd that nobody cares. I know there are others that also had to do that internal debate, and were putting off the answer as long as they could.
I know a lot of people outside the US (but not all) have much longer periods of time per year that their employers allow them to go on holiday, but in the US, unless you’re with an employer for at least 5 years, most people don’t have more than a total of 10 days off beyond the state/federal holidays that are commonly given, and also, a lot of people for many reasons often get stuck having to use their vacation days for personal days, due to how things are done there (sick days are often much rarer than paid holidays in the US) such that there are people that can’t take that time off without a rather lengthy period of warning, if nothing else, to accrue days they can use. WalterCon 2005 and 2006 (both I’ve been able to attend, while being sleep-deprived from travel arrangements) were also rather short notice. I believe Michael Phipps will have a hard time rebuilding trust that this won’t happen again, which may end up killing off Haiku, Inc. planned WalterCons, and perhaps leading to more FalterCons.
We shall see.
Those of you who bought tickets via credit card (which should be 100%) can easily do a charge back. Regardless of what they say about ‘nonrefundable’. If they don’t deliver the goods advertised (in this case: not holding the conference on the date/time agreed) they will lose the charge back. A couple years ago i booked airline tickets via a popular travel site. When i got to the terminal, I could not check into my flight. The excuse given was the flight was booked and that I’d have to be transferred to a later flight (in this case: The next day.) I promptly issued a charge back, even though the tickets were listed as ‘nonrefundable’. I received my money.
That depends on the issuer. One of my credit cards is a premium card geared to travel, and I can get a credit for any flight delayed more than 4 hours. That’s the credit card’s policy, they don’t recoup that from the airline. It’s effectively a self-funded group insurance policy offset by the ridiculous fees I and all the other suckers pay annually for the card.
Generally speaking, credit card companies will not provide a chargeback for a delayed/bumped/re-scheduled flight; if yours did then it’s likely the airline never bothered responding to their claim (which is a standard procedure for charge backs), or you’ve got a gold card or similar specialty card. As far as the credit card companies are concerned, you’re purchasing a service, and that service is to get you from point A to point B. The airlines have enough claims and waivers in their terms and conditions that they are pretty much absolved of liability for being unable to provide a seat, unless they are unable to get you to your intended destination in a reasonable amount of time.
Discount/web fares tickets are a great way to save money when flying, but you get what you pay for, and the airlines are pretty blunt about the terms and conditions, though few people bother reading the fine print. When the airlines sell seats at that rate, they’re gambling on the statistical probability that customers purchasing higher fare (and higher priority) tickets won’t actually be boarding that flight. Customers in turn are gambling on the airlines ability to gamble, by taking that cheap rate, although many don’t necessarily realize it at the time. When the airlines don’t gamble correctly, flights get oversold. Sucks, but it happens. A lot.
Point being, chargebacks against an airline because you couldn’t board your flight are generally overriden unless the airline failed to provide a reasonable alternative to get you to your destination. So you either got lucky or were simply exercising a privilege your card provides, but I wouldn’t promote it as a failsafe method. If you’re going to fly discount, you’re taking chances the same way as when you rent a car without the add-on insurance (another liability covered by some credit cards that many people don’t realize). You’ll probably be ok, but you’ll curse up a storm and have no recourse if something goes wrong.
“Pardon me if I sound irritated, but I find it aggravating that the same general group of people (OSAlert readers, that is) are very vocal to praise Haiku’s achievements, but also very quickly backstab”
Noone backstabbed noone. Just airing frustration over a moronically organized OSS project.
Not every OSS project is backed by a big company (like RedHat, IBM, etc). So please do not get frustrated if people’s voluntary efforts are not professional enough.
I’m sorry, but if anyone here got stabbed in the back, it’s the people (like myself) that spent hundreds of dollars on airfare only to have the rug pulled out from under us without any warning.
Just because Haiku, Inc. administrates an OSS project doesn’t mean they get a free ride when it comes to piss-poor business sense.
Granted, they’re still not anywhere near the level of YellowTab.
Keep in mind Haiku, Inc. != Haiku. Haiku can live just fine without Haiku, Inc… but it sure helps to have the business entity around. The problem is, they need to start -acting- like a business entity, instead of someones hobby.
I’ve wasted a substantial amount of real, hard-earned money only to have the primary reason for a trip to be canceled without warning, and now I’m a bit pissed off.
Just for the sake of (hopefully) quelling misunderstandings, WalterCon itself is completely refundable and, if you read the post on the main site, it even says so. The problem is that people got shafted on nonrefundable plane tickets as mikesum32 pointed out. Traditionally, Michael Phipps was responsible for organizing WalterCon. As much as he is a dear friend to me, I personally think he has been so busy that it was put off. This year’s disaster won’t happen again — I’m organizing next year’s WC and I’m already planning it.
The day after the announcement, I posted a full explanation of what happened. You can read it at http://www.freelists.org/archives/openbeos/07-2007/msg00175.html
The situation was not nearly as simple as you might think, so please save yourself some possible embarrassment before posting.
Pardon me if I sound irritated, but I find it aggravating that the same general group of people (OSAlert readers, that is) are very vocal to praise Haiku’s achievements, but also very quickly backstab (at +4 to Hit, for double damage) the project when something goes wrong. We don’t do this because we’re paid for it. There are a variety of reasons why, but one of the main ones is that we want to help people by creating an OS which is a viable alternative to Windows. We’re human, guys, so cut us all some slack. You’d probably appreciate a little if the tables were turned.
Just so it’s absolutely clear – We the FalterCon organizers and attendees praise DarkWyrm’s efforts and attempts at pulling off WalterCon this year despite his late start.
I’m personally a little frustrated that the “cry for help” didn’t come prior to the cancellation announcement, as I think the rest Haiku community *could* have stepped up and made it work.
This to me was simply a communication problem, and I think maybe some things could be learned from this.
FalterCon is not intended as a public stab at Haiku or WalterCon – the name is a joke – to make light of the situation – and we truly mean no serious disrespect.
We believe there will be some positive outcome from FalterCon. At this point, we’ll take the opportunity to reach out to some Linux users and show them an alternative (HA! an alternative to an alternative OS…)
Ditto. I’ve been in contact with DarkWyrm, and he’s been a great help to us in providing us with info for FalterCon. Thanks Wyrm!
The number of balls in the air with WC 2007 were more than could be caught by the amount of people (one) working on WC 2007, and those people have my sincere condolences. I know this kind of thing isn’t easy to plan, or pull off, which is why it needs to be started sooner (like DarkWyrm is doing for next year already).
Hopefully this never happens again, but sadly, until then — people are going to be skeptical.
And it’s not like people didn’t try to warn Haiku, Inc. months ago that this kind of thing would likely happen if they didn’t get their ducks in a row.
The problem is that people got shafted on nonrefundable plane tickets as mikesum32 pointed out.
That’s an error on my end. In Dutch, the word ‘ticket’ is solely used in the context of an ‘airplane ticket’ – for other uses, such as tickets to the cinema or theater, we have a dedicated word (‘kaartje’, which means ‘little card’ in English). So when I read ‘ticket’, I immediately thought of airplane tickets, not realising the word is used in English for all sorts of different tickets.
I added ‘airplane’ to the teaser to clarify.
Gotta love the EASL (english as a second language) effect.
Thom, thanks for the clarification, and now you know for the future!
Pardon me if I sound irritated, but I find it aggravating that the same general group of people (OSAlert readers, that is) are very vocal to praise Haiku’s achievements, but also very quickly backstab (at +4 to Hit, for double damage) the project when something goes wrong.
No matter how annoying that might be, it’s human nature to do so. At OSAlert, it’s the same thing; we rarely get an email or a comment stating we’re doing a fine job, but as soon as something goes wrong, no matter how minor, our inboxes and comments sections are flooded with complaints. If you’re in the game of providing people with a service or product, free or not, you’ll have to learn to deal with complaints.
Haiku has seen mostly good times, but now that this went wrong, people complain – and rightly so. Just remember that people complain – because they care.
Haiku has a lot to prove of itself before being allowed to make mistakes and get away with them.
So far Haiku is in the same boat as Minix; interesting developments but minuscule community in comparison to even GNU/Linux, let alone Windows.
As far as non-transferable tickets to a canceled conference are concerned, no project can avoid the complaints of angry attendants.
Edited 2007-07-31 08:00
“So far Haiku is in the same boat as Minix; interesting developments but minuscule community in comparison to even GNU/Linux, let alone Windows.”
Don’t EVEN compare Haiku to Minix. They’re NOTHING alike, developmentally. Go to Haiku CIA and look at the numerous commits that Haiku gets on a daily basis. You’re lucky if you see even one or two “news” blurbs about Minix on their own home page.
I respect Minix 3 for what it is and may be trying to be, but it sure as blazes ain’t no Haiku.
I may complain about one particular problem with Haiku, but that’s because want it so badly. It will come eventually. But with Minix… I’d probably never see it.
Unless there’s more Minix 3 development going on than they let on, I’m pretty sure they only have a couple people coding a few bits and bytes here and there, whenever the mood hits them.
Remember… Minix was and will likely always be considered… an EDUCATIONAL OS. Mainly to *learn* from, not to ever be truly *used*.
Well what I meant was Minix and Haiku are similar in that they are minuscule in usage and perhaps even irrelevant to the rest of the world at this point.
They are both “toy” OSes until they prove themselves otherwise over a long period of time.
“Well what I meant was Minix and Haiku are similar in that they are minuscule in usage and perhaps even irrelevant to the rest of the world at this point.”
I take the potential of Haiku a LOT more seriously than I do whatever potential Minix has/had. Haiku is not a “variation” of Unix. Minix is. Just like Linux. It’s just that Linux wasn’t constrained by the whole “educational” limiting label, like Minix was. As I see it, there was, fundamentally, NOTHING preventing Minix from being the Linux that Linux became. Nothing except a label and people’s dogmatic desire to protect that label!
“They are both “toy” OSes until they prove themselves otherwise over a long period of time.”
To me, a “toy” OS is one that is written by one person (maybe two to four, at most), which is not designed to ever become something more than a learning exercise. Minix *could*, in a sense, be classified as that, even though I can’t really justify that title, given that it holds such potential beyond it’s design.
” As I see it, there was, fundamentally, NOTHING preventing Minix from being the Linux that Linux became.”
Except maybe a fundamentally piss poor micro-kernel implementation.
“To me, a “toy” OS is one that is written by one person (maybe two to four, at most),”
Another definition of a toy OS is one that doesn’t work reliably.
Edited 2007-07-31 22:21
“Except maybe a fundamentally piss poor micro-kernel implementation.”
I do not believe that Minix is/was limited by it’s kernel design, but the “educational” title. Because Tanenbaum said “this is for educational purposes”, all that have worked on Minix, have heeded that edict, as though it were gospel. As though it could NEVER go beyond just an “educational OS”.
And if, by chance, the kernel WAS the hindering point, maybe it’s BECAUSE it was always worked on from an “educational” perspective and little beyond that…
“Another definition of a toy OS is one that doesn’t work reliably.”
Depends upon how you look at it. A “toy” is something you play with. A toy can never be more than what it is. It was DESIGNED to be what it is. A toy never evolves or grows into anything better or different. Can a pogostick ever be made into a bike? Or a Yo-Yo into a Basketball?
HaikuOS, by that concise definition, is *NOT* a toy. It is a “work in progress”. It is improving. It is changing. And it will arrive in R1 form… eventually (let’s strike up the Jeopardy theme, while we wait!).
Although the BeOS/Haiku scane is one of the biggers in the range after the Windows/Linux/MacOS trio, I am not sure it has reached (yet) the “critical mass”.
The Haiku project is indeed one of those that have a chance to end up as a success, but they must find the ways to come out with something that makes the project even more attractive.
Wishing the bests anyway.
(otherwise, if it had been planned for Europe, I was on the list of money loosing guys)
[q](otherwise, if it had been planned for Europe, I was on the list of money loosing guys)
Maybe not. Most of the Haiku developers/users are in Europe, and August is the typical European holiday month, so I guess it could get enough people to attend.
The only problem could be the late announce.
Since you guys have so much money to throw about then why don’t you contribute to the programmers so they can get Haiku (BeOS) developed before, let’s say, 2025?
“Not every OSS project is backed by a big company (like RedHat, IBM, etc). So please do not get frustrated if people’s voluntary efforts are not professional enough.”
Even when that means loosing money?
Size is no excuse. Some of the most successful, professional, best run companies I’ve worked with have had fewer than 25 employees, and started in their mother-in-laws basement.
In fact, BECAUSE you’re small, you have to be that much -better- than the people who are bigger than you. Claiming that because you’re small you’re allowed to fall flat on your face only highlights how out of touch with business reality Haiku, Inc. is.
Things developers should not be allowed to do:
1. Design an application – things will be complicated, very hierarchical, etc. The average user will be lost.
2. Manage the finances – new AlienWare laptops for everybody!
3. Interface with the public – they’ll bite your hand off. Please don’t feed the programmers, madam.
4. Run the organization – or you won’t be able to call it an organization!
p.s. I’m a developer!
I’m hoping you mean don’t let them design the UI
I think there may be other reasons no one signed up. WC 2004 Columbus,OH was $30(just conference/shirt), $75(con + double occupancy), $120(con/shirt + private room); 2005 Las Vegas was $35(con/shirt) + $75(room)=$110; WC 2006 Orlando, FL was $125 (no option for room); WC 2007 $145.
I think it would be help people make up their mind if the cheapest just-shirt-option was reinstated. It would make the decision to attend easier if they could get the most out the money they spend. Maybe they want to spend money at a cheaper hotel. I think there should be an option for locals who might stop by for a look-see, of course I don’t think this applies much outside of Silicon Valley. More money upfront leads to more hesitation, in my humble opinion.
I also feel that Waltercon will never be as popular as the first one, at least until R1 is released. The first one was an exploration of why Haiku is here and where it’s going, and also a place to meet the people involved. Perhaps some people were looking for a place to belong. They had to make a choice to make the next year. Some people got on each others nerves, realized Haiku wasn’t the answer to their problems, or they were just not impressed. There are also money, travel distance, and time-off-work issues as well. I think that covers most of the reasons. You’d have to ask the people who went to the first one, but not to any of the others why. Takeshi Takasugi flew in from Japan for the first one. Deej went to the first one, and he’s still involved with BeGroovey, but he didn’t go to any other ones. Insight would help fix problems. Bryan Varner had a few suggestions/complaints a while back on his blog.
Another thing, we should try to network with other communities(LUGs anyone ?) and maybe send some polite invitations to former attendants in the area. FalterCon has managed to whip up more attendees than Waltercon did. What does that say ? I think it says proactivity is more effective than passivity. Get up and make it happen, just don’t let it randomly unfold. Of course the promise of free food is alway an effective lure.
Hopefully, the number of people should be closer to the Las Vegas numbers in the future, if the city Chamber of Commerce and gmail don’t conspire against us, and of course if we can fix any and all failings.
I also made up a fake shirt for FalterCon, although I might change it some more later. I don’t think anyone likes it, but there is always the possibility people might like the redesigned one even less.
This was done in irreverence, not anger.
http://img2.freeimagehosting.net/uploads/886013ca07.jpg
I’ll try my best to make sure the next Waltercon will go well, and I’ll make lemonade out the lemon that is waltercon 07. Faltercon is going to rock. I’ll be surrounded by Linux nerds.
I was getting Deej confused with Daat and was going to make a correction here, but noticed I was right the first time. Time for bed.
Edited 2007-07-30 14:37
There are a few things that Daat left out that would make this situation more clear.
1) A grand total of *4* people signed up. 4. While I very sincerely regret ANYONE losing money on plane tickets, and I have personally apologized to most of the people involved, this isn’t as widespread as the article makes it sound.
2) The survey we ran this year asking the community about what they were looking for in a WalterCon indicated that what we proposed to do for 2007 was the RIGHT thing to do. The location, month, etc were all determined by the community.
3) NOT ONE PERSON has written to me, despite my invitation, either publicly or privately to tell me that they would have come if they had known earlier. Not one. So can we *PLEASE* bury this idea? It is not based in any reality at all, only people’s beliefs and guesses.
4) Haiku, as someone pointed out in the ico forums, is COMMUNITY RUN. If you want to make it better, help out. Complaining in forums only HURTS the project. It doesn’t HELP anyone. We have repeatedly asked for help in many areas.
5) We APPLIED for Google Summer of Code last year. We were not accepted. The referenced article would have you believe that we didn’t bother. Absolutely not the case.
6) Haiku’s board of directors was never a question that was ever brought up. If *ANYONE* would have asked, they would have been told. In fact, New York law says that they have to be told.
7) As far as bounties go, the posted data is part of the story, but far from all. There have been a number of private communications between Haiku and Karl. Yes, my machine crashed and I lost the email that I needed to respond to. I couldn’t help that and I did post to the mailing list, if I remember correctly, that that had happened. I regret the fact that Karl and I had that issue, but that was almost a year ago. As far as bounties go, we proposed a much reduced scope for bounties and I asked Karl to do some footwork to make it happen. He indicated to me that he couldn’t commit to that footwork and it was left at that. This is, as I said, a community run project. We are far from over run with people who want to do the work that needs to be done.
8) There is a perception that we don’t communicate with the community. We are more open than ANY OTHER OSS project of any size. The mailing lists are open. The admin meetings are summarized and posted. SVN logs are open and are now even on our website. There is *NOTHING* that we discuss that the community doesn’t know about. WE CAN’T BE ANY MORE OPEN. There is nothing we hide.
Please, can we stop wasting time and effort on fighting with each other? If you want to know, ask. If you want it better, help. Don’t make blog postings that complain about people who are doing what you won’t. That doesn’t help Haiku. It doesn’t help anyone. Haiku needs more good people who want to help. Several people, recently, have joined and are helping. It can be done.
You might have missed this. http://www.freelists.org/archives/openbeos/05-2007/msg00224.html
and this
http://www.freelists.org/archives/openbeos/04-2007/msg00517.html
and this one too
http://www.freelists.org/archives/openbeos/05-2007/msg00185.html
Please elucidate.
*Edit*
6) Haiku’s board of directors was never a question that was ever brought up. If *ANYONE* would have asked, they would have been told. In fact, New York law says that they have to be told.
You are the board of directors ?
http://www.freelists.org/archives/openbeos/04-2007/msg00514.html
– Koki mentioned Haiku Inc has a board of directors, three including yourself.
Would you mind saying who these other people are?
Me. Everyone else has become inactive. We had some internal discussions about adding more people, but it really didn’t go anywhere. The biggest issue was/is that I would like a BOD that is fairly balanced. Right now, we have very few contributors who are not devs. I love devs. I am a dev, at least at my paying job. But I think that we need other (business/marcom, legal, etc) representation. OTOH, there is a lot of concern around inviting people to Haiku who are new, or at least, don’t have a lot of knowledge and vested interest.
Busted. Here is your chance to make it right.
Edited 2007-07-30 16:40
8) There is a perception that we don’t communicate with the community. We are more open than ANY OTHER OSS project of any size. The mailing lists are open. The admin meetings are summarized and posted. SVN logs are open and are now even on our website. There is *NOTHING* that we discuss that the community doesn’t know about. WE CAN’T BE ANY MORE OPEN. There is nothing we hide.
It’s not about hiding, it’s about engaging. Do you see the difference?
The example I provide above is the lack of community involvement in the decision to cancel WalterCon… If the community had been alerted that it was going to fail, then the *community* at that point would have the onus of making sure it didn’t.
Instead, the decision to cancel was made and announced publicly after the fact. What good does that really do other than to notify people to stop registering?
If someone is going to divorce their spouse, “openness” would be talking about it with them BEFORE you file legally. At least that way there’s a possibility of resolution without legal proceedings, and if not – you proceed as planned.
Turning a bad situation into good PR is also possible… (look around, you see this all the time) This would have been the right opportunity to alert the community about the problem, show them that the US does not necessarily have a very strong Haiku community as it does in Europe, and CHALLENGE them to make it better.
Just like you, we’re all busy working people too (well, most of us) – but we’ll step up if necessary and find a way as a community to make positive things happen.
Anyhow, I’m out of time and must go to work now.
Well, Michael, only 4 had signed up officially before it was cancelled, but there’s a major problem with that:
The signup deadline was after it was cancelled.
I was intending on going, but had to judge my current situation as close to the cutoff time as possible. I’d wager there were an unknown number of others that were doing the same thing, for the same reasons: it isn’t always feasible to take off the time, get the money, etc. with that short of notice, even if they had intended on going since last year.
I hope that these lessons are learned and applied in the future:
1. Put absolute deadlines into everything, as to when people must be signed up on the participants going, so Haiku, Inc. has a known quantity.
2. These deadlines must also be communicated concretely to all venues, etc. involved in the whole thing, and must be part of the deal: they, too, need these deadlines so they can plan: it’s only good business.
3. Make it very clear that a renegging venue or one that even hints at screwing Haiku, Inc. and participants over will have their names dragged out in public for their bad business dealings, and if it comes to that, do it without reservations (pun may/may not be intended!). It sounds very much like that’s what has happened here: they figured that with the short timeframe left, they could put the screws on Haiku, Inc. to commit or they would take their chances with making what they feel are more profitable arrangements by screwing over Haiku, Inc.
4. Those deadlines really need to be > 2 weeks in advance of the event, not < 2 weeks like it was, because if nothing else, making flight reservations and room reservations with less warning than that gets much more expensive, and besides, wouldn’t it be nice to have more lead time to ensure the correct number of t-shirts were printed so there aren’t a huge number of unsellable leftovers?
I would suggest that you contact the businesses that decided to screw over Haiku, and inform them that you will put in a very public place online how they’ve done business with all the gory details, do as you stated, and leave that as proof in the future for all others that you will publish how they do business: if they’re doing their jobs legally and morally and correctly, they should be overjoyed with the publicity: if they’re doing things in a shady manner and not honoring their promises, they should be afraid of that. As long as all details are correct, they can’t possibly sue.
They have been taking notes from ZetaOS. Never, ever tell anyone in your community what is going on. They may realize that you are not as smart and organized as you want them to believe.
MikeS – as I said, the others are inactive. Mike Wilbur and Tyler Dauwalder are the two still on record, if you really want to know.
Urias – we didn’t get a lot of warning from the hotel – they gave us a couple of days to make a decision. We couldn’t have done much better.
And I will (again) point out that not one person has come forward and said that they would have come if we had announced earlier. Nor did anyone complain about the location or the price. In fact, the survey showed that most people don’t care about the price, within reason.
Oh – and most of the other WalterCons were planned by one person (myself, the first, Czeslaw the second, myself the third, DW this last one). DW explained in depth the problems that we had this year. We couldn’t have anticipated those problems, nor is there *ANY* evidence that starting earlier would have helped.
Urias – we didn’t get a lot of warning from the hotel – they gave us a couple of days to make a decision. We couldn’t have done much better.
Ok, yes – I did get a whiff of emergency, reactionary measures here. I’m still not convinced the strategy chosen was the best – but what’s done is done. I am still trying to offer some constructive criticism to help avert similar situations in the future.
And I will (again) point out that not one person has come forward and said that they would have come if we had announced earlier. Nor did anyone complain about the location or the price. In fact, the survey showed that most people don’t care about the price, within reason.
I will come forward and say that if the date and location had firmly been announced publicly at least 3 weeks sooner, I would probably have signed up firmly. As it was, I had to plan my August vacation time off, and had to plan it around some local events in my county (our county fair occurs the week of August 8-12, and I’m sure you know how important county fairs are to children when you live in a rural area
. ~3 weeks sooner and I would have had a “clear schedule” as far as I was concerned (only to have my wife complain to me later).
If I had more advanced notice, I would have been able to juggle things a little better and make sure my family and employment plans were taken into account. As it was, I was still trying to make compromises in order to go, and if I had known it was in danger of cancellation, I would have immediately registered. As JT says, there was still time left before the registration due date.
At this point, I’ll be going to FalterCon only on Saturday the 11th, so I can spend time with my family again on Sunday. This actually works out better for me, and made the decision to attend FalterCon much easier. It also helps that it won’t cost me much more than the gas it takes to get there and home, as well as time and energy spent in preparations.
You can blame all sorts of unfortunate issues that caused the end result, but more advanced notice would have done nothing but help in this case.
I’m glad to hear this problem will be resolved next year, and I will try and help make sure it doesn’t if there’s anything I can do (even though I realize it’s probably not going to be in my local area again soon).
Michael, take a deep breath, and read what your own people are saying.
DarkWyrm:
(http://www.freelists.org/archives/openbeos/07-2007/msg00159.html)
The perceived reason for this would be late announcement combined with the community being spread across the globe.
(http://www.freelists.org/archives/openbeos/07-2007/msg00175.html)
Michael and I decided to drag our heels on the contracts until after
the registration period officially ended. The guy from the Conference
Center e-mailed me Friday and told me that the contract / payment were
due 7/13 but we had until 7/25 before they’d consider it canceled
Timeline of the WalterCon 2007 march to failure:
7/3/2007 – Announced
http://www.freelists.org/archives/openbeos/07-2007/msg00008.html
Contracts are due 10 days later. Presenters and attendees have about 39 days to get ready.
7/11/2007 – Registration now Open
http://www.freelists.org/archives/openbeos/07-2007/msg00036.html
Contracts are due in two days. According to the official WalterCon site, registration closes in 19 days.
7/25/2007 – WalterCon is canceled.
http://www.freelists.org/archives/openbeos/07-2007/msg00159.html
Five days before registration closes.
It is later disclosed that contracts for the convention center are overdue by 12 days. Contracts for the hotel have never been received to be signed and returned.
http://www.freelists.org/archives/openbeos/07-2007/msg00175.html
Your indignant claims that time was not (and has not ever been) an issue and WalterCon can be safely and happily thrown together in 49 days have no basis in reality. If these claims reflected reality, I’d be spending August 11th and 12th in a convention center and sleeping in a Four Points Sheraton the night of the 11th.
Stop ignoring reality. Just closing your eyes doesn’t mean it’s not there.
Edited 2007-07-31 19:03
“Stop ignoring reality. Just closing your eyes doesn’t mean it’s not there.”
Same goes for fundamental and structural flaws.
Why are you acting like I’ve attacked you?
My SMP box works with Haiku. I cannot stress enough that this has not always been the case. The exact incompatibilities were ironed out on my hardware. I got lucky, Axel has a similar machine.
Predictor, I agree that disabling SMP early on was asking for problems. I raised these concerns on #haiku once upon a time, and geist very vocally agreed. I think the general consensus was that it was a dangerous move, but was made for the sake of functional progress in a few key subsystems. SMP has been re-enabled for over a year now, and works well on some systems. On others, it chokes.
I agree that continuing to develop on emulators rather than real hardware will also hinder how well it runs on real hardware, especially in compatibility.
BUT I can also respect those that use virtulization and emulation as a means to speed development along. I understand and respect that using a controlled environment to eliminate hardware discrepancies leads to a more complete system sooner than developing on real hardware. Once the core system is in a known-good, working state, we can worry about hardware inconsistencies and drivers. The primary goal is to have an OS that is stable and can host itself. Once that is accomplished, major systems don’t spend as much time in flux, which makes it easier to focus on drivers, HW compatibility, and minor revisions. It will also give us a good, working codebase to make as you put it, ’64 bit safe’.
Of course, all of this takes time, patience, and developers willing to donate time and energy to the project.
In an attempt to give this a positive spin, I have a couple of requests for Mr. Phipps that could help maximise the Haiku PR effect of FalterCon.
* Would you post information about FalterCon at the top of the Haiku website front page with a link to http://myhaiku.org/faltercon, so that more people get to know that we will be there, and as a public endorsement for this community effort?
* Would you provide us with Haiku t-shirts to wear at the event?
If mphipps would agree to koki’s suggestions and koki would consider taking up his marketing position with Haiku again and everyone learning from what happened, letting bygones be bygones, I promise to swiftly donate $100 to Haiku or may the gods strike me down.
From the WC07 fiasco and Mr. Phipps response to the community here on OSAlert, it’s pretty easy to tell that nothing has really changed at the heart of the project when it comes to non-development activities. I *have* learned my lesson, and would not want to go back to the same status quo that lead to my parting from Haiku.
So, thanks, but no thanks. I still have fun doing stuff with the community from time to time though. FalterCon is one of those things that I find fun to do.
Having said that, I would still encourage you to donate those $100 to Haiku anyway. IMHO, a good place to put your money would be one of the ongoing bounties at http://www.haikuware.com. Personally, I like the Webkit Port bounty: it has been assigned to a dev (Ryan Leavengood), and the port has already started. You can even find a “roadmap” of the project at an article here:
http://joomla.iscomputeron.com/index.php?option=com_content&task=vi…
Edited 2007-07-31 02:12
Hm, Haiku is getting a lot of bad vibes. WalterCon would’ve had FOUR people. Haiku doesn’t boot on most of my REAL machines (Haiku developers: please stop using those f–king emulators all day).
Maybe dropp the “board of directors” and start a proper meritocracy, no?
Edited 2007-07-30 22:25
Hm, Haiku is getting a lot of bad vibes. WalterCon would’ve had FOUR people. Haiku doesn’t boot on most of my REAL machines (Haiku developers: please stop using those f–king emulators all day).
You repeatedly appear to hold a grudge against Haiku because you can’t make it run on your hardware – this is probably the 3rd such time I’ve seen you use that as a platform to abuse the project…
Emulators are a known good way to test a system in a known environment. The world of PC hardware is a maze of disaster that nobody has time to deal with yet while they’re trying to solidify the core OS kernel and other such items. Why should Haiku care about some specific piece of hardware now when they don’t even have an OS that is finished?
Your attitude is pretty much exactly why Haiku hasn’t released yet to the public even as an alpha.
You are making dangerous assumptions: I intended on being there (I was at the past two), but Michael Phipps also made the dangerous assumption that you are, that those that signed up before it was cancelled were going to be the only people there. Facts are that there’s no way to be certain how many would have signed up before the deadline, because it was cancelled before the official deadline, which reasonable people would believe to mean that they had until that time to commit and get the required entry fees paid to Haiku, Inc. which was running on way too short of a notice for people: after all, last year’s happened in November, and even though it was suggested it’d be around August this time, for the longest time, it didn’t look like it would happen in August, or at any predictable time at all. Sadly, that appears to be the case. You can’t make plans with such short notice when you aren’t absolutely confident it will happen, especially when you have the last year’s results to demonstrate that everything was likely to be up in the air.
And as Urias has said, well, your hardware, however broad-based and generic you might think it is, does not represent everything available, and BeOS itself never ran on everything under the sun, and heck, Linux still doesn’t have hardware support that allows it to run on everything under the sun (though still better than BeOS/Zeta ever did overall) so if you really want to be useful, start working towards figuring out why it doesn’t boot on several of your real machines, and become a tester. This will likely require using a serial terminal on the working system, and getting debugger output and submitting it to those that have a clue what’s going on, or, if you have a clue as to what’s going on, trace through it yourself. Or, another alternative that may be far more valuable (if you’re committed to getting things to run, and are willing to do more than put out complaints) ship your non-working systems to the team so they have real live hardware to test on! Nothing beats real hardware for testing the quirks that exist, but if they don’t have it, they can only test what they have available. If I were really that upset about my machine not booting, I’d find a way to help make it work, but… all of my PC hardware is BeOS and Haiku compatible.
“You repeatedly appear to hold a grudge against Haiku because you can’t make it run on your hardware”
Well, according to #haiku, not beeing able to boot is a common occurence these days. SMP (which is totally common now) is simply broken.
I have no grudge against Haiku beyond the fact that developers seems to be incompetent in certain areas (64-bit cleanliness, SMP and quite frankly testing on proper hardware)
“Your attitude is pretty much exactly why Haiku hasn’t released yet to the public even as an alpha.”
Who’s the “public”? I’m usually fully capable of getting hobby OS’s to boot. Not so with Haiku.
SMP is broken and doesn’t boot on a multitude of machines. Yet, you guys have a very professional looking web site with a BIG FAT image download button and are looking for opportuneties to “sell” the project. Focus on major remaining technical issues first, thank you very much.
Just a friendly suggestion, really.
Edited 2007-07-31 06:49
Well, according to #haiku, not beeing able to boot is a common occurence these days. SMP (which is totally common now) is simply broken.
You’re actually referring to some statements that I (and a couple others) said. I would be glad to dredge up the logs to show that you are blowing some those statements way out of proportion. You came into the channel and several tried to be helpful, giving you some things to try – but it’s really one of those things that requires a kernel hacker to track down and investigate on real hardware.
What you’re leaving out is that several people in the same channel claim that it works fine, and Axel has already responded to you directly stating that all of his SMP machines run Haiku fine:
http://osnews.com/permalink.php?news_id=18199&comment_id=252886
Furthermore, when Haiku was started, x86-64 didn’t exist yet… and it’s not a goal of Haiku R1 to be a 64-bit OS.
In any case, nobody’s “selling the project” in the sense you’re insinuating. In almost all cases, the project is looking for help and contributors – something you’re not really doing very effectively at this point. It’s hard to ask for help if you can’t attract people – the “big fat image download button” is there to help people see it hands-on.
You act like every other 6-year-old OS out there written by a handful of people on an almost-exclusively volunteer basis is capable of running flawlessly on all 64-bit/SMP hardware…
I’m not trying to be an “apologist” here either – I’m just trying to inject some reality into your arguments.
(update, added link to axeld’s post)
Edited 2007-07-31 07:21
“SMP is broken and doesn’t boot on a multitude of machines. Yet, you guys have a very professional looking web site with a BIG FAT image download button and are looking for opportuneties to “sell” the project. Focus on major remaining technical issues first, thank you very much.”
And *I’m* still waiting for Haiku to be able to copy files correctly, without eating up all my RAM and dropping to KDL, if I go one byte over…
We all know that already.
SMP does work with the machines of the current Haiku developers. According to the bug tracker, there are no known SMP issues. Did you file a bug report? The bug tracker is located at http://dev.haiku-os.org/
Going to #haiku and complaining is as effective as telling it your mom. She probably doesn’t know what you are talking about, and can’t help you.
regards
Marcus
“SMP does work with the machines of the current Haiku developers.”
That doesn’t tell us anything. Of course it works with the machines of the developers. That doesn’t change the fact that SMP is broken.
“According to the bug tracker, there are no known SMP issues.”
That’s a big fat lie.
http://dev.haiku-os.org/ticket/1153
In addition, there are CLOSED tickets on SMP problems in there, because the submitter didn’t provide enough info (or they were unable to reproduce – it doesn’t say!)
Such as http://dev.haiku-os.org/ticket/34
“”According to the bug tracker, there are no known SMP issues.”
That’s a big fat lie.
http://dev.haiku-os.org/ticket/1153 ”
Which is not a kernel problem. Haiku has to boot on that pc to be able to run that application, right ?
And that’s only that application which crashed, not the whole OS.
Predictor wrote:
Ticket 1153 is an open issue that might be related to SMP. However it’s completely different from what you have been referring to so far.
Ticket 34 was closed as invalid, because it didn’t contain any describtion.
Ticket 1018 (you didn’t mention that one, I know, but this was a real SMP problem) was closed because the described problem was fixed.
At this time, there is no open ticket about SMP problems as you described. You also indicated that you don’t want to invest the time required to register and report a bug. Don’t expect any of the voluntary Haiku developers to start working on such a basis.
“At this time, there is no open ticket about SMP problems as you described.”
Ignorance is bliss.
Then why aren’t you happy?
They’ve given you the requirements to get such things at least looked at, but you clearly refuse to do anything but whine incessantly where it does no good.
… a strong consensus of the Haiku Community. We need to know who is a developer and who isn’t. We need to know who “watches” Haiku’s development (from the sidelines), yet is interested/able to attend a WalterCon. We need to know who lives where, Country (outside of America) or State (within America). We need to know how far each person is willing to travel and by what mode of transportation. We need to know how much each person is willing to pay (in absolute sum total) to attend a WalterCon. WalterCon should be held at the same place/month/date every year, no if’s, ands, or buts. If the majority of Haiku devs and “watchers” are located in Europe, hold all WalterCons there.
Make everything so absolute from the first, that there are NO misunderstandings, no late signups, no cancelled tickets, etc. Even if such stiff requirements are an inconvenience to some, at least everyone is on the same page, from page 1! And, thus.. no angry Haiku Community members spitting venom over losing a few hundred dollars for a ticket they can’t use now, because WalterCon was canceled.
What we need is someone willing to lead the charge and for the Haiku community to speak up and be heard. Every last person who knows even anything about Haiku should be part of this. Do we really want another “FalterCon” to mock Haiku’s blunder? Is anyone fooled by the name usage? You can say it’s not meant to be an insult, but it’s so obvious, it’s not even funny. If I were one of the ones that was burned, I wouldn’t go, just because of what the name represents! It’s bad enough that some people lost hundreds of dollars… but it’s even worse to create a mockery event of the catastrophy!
As the saying goes: “Do it right or don’t do it at all.”
Do we really want another “FalterCon” to mock Haiku’s blunder? Is anyone fooled by the name usage? You can say it’s not meant to be an insult, but it’s so obvious, it’s not even funny. If I were one of the ones that was burned, I wouldn’t go, just because of what the name represents! It’s bad enough that some people lost hundreds of dollars… but it’s even worse to create a mockery event of the catastrophy
Ha! you must realize the same people that got burned came up with the name…
I’ve never understood why people get so upset about stuff like that. People need to get thicker skin and laugh at life a bit more.
In any case, FalterCon is going to happen, and it’s going to be a positive event. We believe it’s an opportunity to prove that the community can survive mistakes made, and can rise up to the occasion – using this opportunity to “invade” (with permission of course) a Linux gathering and bestow upon them the greatness that is Haiku.
BTW, if anyone should be offended by the name, it would be the guy who coined the name Walter:
<bga> Although I take offense at the FalterCon name. I came up with Walter, after all.
oh wait…damn (actually, I believe he was mostly joking – I hope!)
(oblig wikipedia link: http://en.wikipedia.org/wiki/Walter_The_Operating_System)
Edited 2007-07-31 08:18
“Ha! you must realize the same people that got burned came up with the name…”
Sure, it seems like it’s a “make the best of a bad sitation” type event, but… it seems to mock your own misfortune rather than standing out and saying “We won’t let this happen again!”. I dunno. I know I wouldn’t attend, even if I were burned. But, looking back, it’s easy to see how badly handled this ex-Waltercon was. And all FalterCon does it make a mockery of the fiasco (and the company behind it) rather than making a stand against such bad management of the event.
“I’ve never understood why people get so upset about stuff like that. People need to get thicker skin and laugh at life a bit more.”
I believe in taking pride in a company image and the events they hold. It’s one thing to have things fall apart like they did, but to make a mockery of it and the people who “flubbed up”, by slurring the name of the event… that’s just heineous in my opinion!
If I were the one running Haiku, Inc. and “hosting” WalterCon, no matter how bad of a messup an event turned out to be (and it can happen), I would not stand for people (even the developers) turning the event into a joke. It’s just bad taste, IMO.
If WalterCon’s are meant to promote an image of serious developers discussing serious topics about a serious OS, run by a serious company, then you DON’T make a mockery of the event, when mistakes are made. You only damage EVERYONE’S reputation in the process.
I know MY opinion of Haiku, Inc., as well as all the people involved in WalterCon, has been somewhat damaged by this whole debacle. And even moreso by the mockery of said event.
As I’ve said before… “Do it right or don’t do it at all.” That’s the way EVERYTHING concerning Haiku, Inc., HaikuOS, and WalterCon needs to be handled.
Otherwise, Haiku as a whole (company and OS) will simply be dismissed as a “toy” OS (gah, what an insulting term!) run by a “joke” of company, that can’t get it’s (!!!) together enough to be taken seriously.
If I were Phipps, I’d be pacing a trench in my carpet and fuming about this whole thing. And I’d do EVERYTHING in my power to turn it around! Even if I had to even pay money out of my own pocket, to do so!
Like someone said earlier, just because HaikuOS is an OpenSource project, doesn’t mean Haiku, Inc. shouldn’t be taken seriously as a company. But a disaster like this, sure ain’t helping! And mocking the non-event only makes it worse.
Luposian – you sound like you are proposing a somewhat-totalitarian approach to dealing with this issue.
“They mock us, so they must die!”
“Let’s pay to make this problem disappear!”
Haiku, Inc. could have continue to run the event even if it meant losing money, but they chose not to. As I said previously, what’s done is done – and we’re going to do something else instead.
Here, how bout this: FalterCon is an event where Haiku community members get together and visit Linux gatherings in order to demonstrate Haiku and gain the favor of some Linux users/developers who may not otherwise have even heard of Haiku.
Now is it acceptable? Are you just stuck on the name?
“Luposian – you sound like you are proposing a somewhat-totalitarian approach to dealing with this issue.”
With an Announce-Open-Cancelled (AOC) event timeline of less than 30 days, that goes into entirely new realms of bad event management!
It was so sad (pathetic), I was actually laughing about it! But, then again, I hadn’t just spent a few hundred dollars towards getting to WalterCon.
Events like WalterCon should be held at the same time every year. Same month, same days. That way, “planning ahead” is a no-brainer. either you CAN go or you CAN’T go. This way, nothing like FalterCon *ever* has to be held, to make up… for the screwup!
I propose there be TWO WalterCons… maybe more, if needed. Each one run to a very specific timeframe (month/days), as determined by a vote of the best time(s) members of that area (country) can most likely attend. When the vote is ended, the month/days are calculated to the absolute most balanced date possible. If the majority of people state the same month they’re available to attend, then THAT month is set as THE month WalterCon is held. If the days are kinda mixed around, between the users, then you set as close to the absolute middle of all those days.
It then gets set in stone… “THIS” month on “THESE” days. Either you can or you can’t attend. Those that CAN attend MUST give at least X# of days advance notice that they WILL be coming, barring an absolute emergency. All people are responsible for their own expenses getting to WalterCon and such-n-such expenses, as necessary are stated.
Everything is set in stone. No questions. No waffling. No wasted tickets. No (!!!)ing “FalterCon” to make up for the screwup!
Now, about possibly more than one WalterCon…
People that live in America are fairly close together. People in Canada and Mexico are close enough to America as well… (why do you think G.W. Bush is trying to make all three into one happy North American Union?), so for WalterCon 1, call it WalterCon NA (North America), which covers all three countries (USA, Canada, and Mexico).
For WalterCon 2, let’s assume there are a bunch of users/developers living in Europe. Do those people really want to hop a long plane ride to go all the way to WalterCon NA? I know I sure wouldn’t. for them, you have WalterCon Europe.
And, if there be enough users/developers in countries and areas too far from either of those two locations, then set up a Waltercon there!
Phipps should be (though someone else could take the position) the leader of the NA branch. Bruno or Axel could be the leader of the European branch. Each one is run to the exact same, strict standards.
This prevents long flights, jetlag, higher plane ticket prices, etc. You could set up a big screen hooked up to a bunch of Macs w/ iSights, running iChat and have a big ol’ video conference party or somesuch.
Well… seems like a good idea to me, at least…
I was the first to use the name publicly, and might have been the first to think of it (I had to step out of a boring meeting at work because I was about to bust up laughing when I thought of it) — I doubt that I was the first to think of it, as JonathanThompson is pretty quick with word-games, but hey… Nonetheless, I was the first to use it in public.
And considering this thread in the haiku mailing list Back in April
http://www.freelists.org/archives/openbeos/04-2007/threads.html#004…
It seems pretty obvious to me that attendance, timing, and preparation where the areas that needed to be focused on to make waltercon a success.
Sadly, what I keep hearing from Haiku, Inc. is : We’re only volunteers, so it’s excusable, right?
Wrong.
I may be an outspoken criticizer of the current state of affairs, but I have offered to Haiku every means I have of helping to make sure next year is a smashing success in comparison to anything we’ve done before. I’m hoping they’ll take me up on that. The ball is entirely in their court.
umccullough said:
“I’m not trying to be an “apologist” here either – I’m just trying to inject some reality into your arguments. ”
The reality, which needs not be injected, remains: I have three modern smp machines (core2, amd x2 and a regular two-cpu), none of which haiku boots on. If I choose the non-SMP kernel, then it boots but doesn’t stay up for long. In the latter case, I engaged in a debugging session but the dev gave up.
My consern is that, it seems that by defintion, as long as the dev’s emulators and machines work, then it works and nothing gets done to remedy the problems. At the same time, new stuff gets ported and added at an insane rate. I’m not impressed by KLOC. I’m impressed by stability. Secure code. Helpful and highly skilled developers. Impress me, god damn it.
“What you’re leaving out is that several people in the same channel claim that it works fine”
I doesn’t much matter that axeld’s machines boots fine. That’s what I’d expect, since that’s the machines he tests on. It’s not like I’m not the only one who has reported this problem either…
Now, I have also mentioned on #haiku that I was able to reproduce the hangs in VMWare with a 2-cpu VM on a AMD dual core. How hard can *that* be to reproduce for a kernel developer?
All I’m saying is, at this stage, Haiku needs more test hardware, not necessarily more people.
Yes, I’m really eager to get Haiku up and running
My consern is that, it seems that by defintion, as long as the dev’s emulators and machines work, then it works and nothing gets done to remedy the problems.
You do realise that hardware costs money, right? Axel and the others don’t have a pool to fish in whenever someone complains about incompatible hardware. Sure, it sucks that it doesn’t run on your machines, but what do you expect from a pre-alpha operating system? Operating systems like SkyOS and Syllable are supposedly much further ahead than Haiku, yet they crap out regularly too when it comes to booting.
Heck, not even Linux boots without wizardry on my bog-standard Athlon XP machine.
At the same time, new stuff gets ported and added at an insane rate. I’m not impressed by KLOC. I’m impressed by stability. Secure code. Helpful and highly skilled developers. Impress me, god damn it.
Stability in a pre-alpha state? Do you have ANY idea about software development AT ALL?
Now, I have also mentioned on #haiku that I was able to reproduce the hangs in VMWare with a 2-cpu VM on a AMD dual core. How hard can *that* be to reproduce for a kernel developer?
#haiku is the place to get help, not the place to file bug reports and get them fixed. You do that on trac and the m-l. #haiku Is a place to have some general talk about Haiku, ask some general questions, or to cheer (and hide your sheep) when DaaT enters.
You are expecting the wrong things from Haiku at this stage. Haiku is not feature complete, it’s horribly unstable, it’s in a pre-alpha state, and nobody ever claimed anything otherwise.
“You do realise that hardware costs money, right?”
That’s a valid point, and yes I’m aware of it. That doesn’t change the fact that they NEED hardware more than people at this stage, which is all I said.
This is a point you make only because you think I’m trolling. I’m not – I wish Haiku the very best, but I’m dubious:
“Stability in a pre-alpha state? Do you have ANY idea about software development AT ALL?”
Yes. In fact, I think I could teach the Haiku team a thing or two about proper development processes. Or they could peek to certain other OSS projects.
“You are expecting the wrong things from Haiku at this stage. Haiku is not feature complete, it’s horribly unstable, it’s in a pre-alpha state, and nobody ever claimed anything otherwise.”
Thom, look. I don’t expect stability from “pre-alpha” software. However, I would expect efforts to move in the direction of stability.
That’s not what I see in Haiku. I see efforts in adding drivers, libraries and various pieces of sugar. Haiku grows insanely fast, but is also getting increasingly unstable (I have used Newos [the kernel in Haiku] and Haiku itself on and off for a while, and things were not this bad early on; I blame that on lack of focus… actually, the Newos part seems to be the only really good part these days)
Fixing fundamental problems after the fact is not easy. The source that is added NOW is routinely not a) SMP clean, b) 64-bit clean and c) reviewed securitywise [like extensive use of unsafe C functions that not even windoze use these days – I’ve been told BeOS was very good at this part].
Anyway, feel free to have the last word. I’ve aired my concerns and will say no more.
Thom, look. I don’t expect stability from “pre-alpha” software. However, I would expect efforts to move in the direction of stability.
You claim to know “a thing or two” about software development, but you’re not really showing that here. Stability is simply NOT A CONCERN at this point, because Haiku is in a PRE-ALPHA state. Currently, the focus is on getting all the r5 features in Haiku, after which stability will be next. You may not like this type of development process, but it’s the way it is.
And as said, nobody ever claimed Haiku is stable, or that stability is currently its primary focus. Your complaints are analogous to a consumer complaining at an electronics store that his TV doesn’t cool his beer.
Fixing fundamental problems after the fact is not easy. The source that is added NOW is routinely not a) SMP clean, b) 64-bit clean and c) reviewed securitywise [like extensive use of unsafe C functions that not even windoze use these days – I’ve been told BeOS was very good at this part].
The only reason you list SMP as a fundamental problem is because your machines are all SMP (I don’t classify dual-core as SMP, but that’s just me, I’m weird). SMP is, by no means, a critical feature in a pre-alpha state – no matter how often you repeat that your machines don’t work. Heck, not even r5 can run on an Athlon properly without special patches, so why should Haiku, a re-implementation of r5, run super-stable on a Core 2 Duo?
Anyway, feel free to have the last word. I’ve aired my concerns and will say no more.
That’s the easy way out. Could you please put this at the beginning of your comments so I don;t waste my precious time?
Thank you.
“Stability is simply NOT A CONCERN at this point, because Haiku is in a PRE-ALPHA state.”
This is funny. As though you can build a house, shoddily, and THEN fix/replace everything after the house is complete.
I have said, many times, that I think that Haiku should be built from the ground up, not built with people going every which way, as they please, figuring, eventually, the OS will “come together” and be complete (R1).
File copying is a VERY basic OS process. It is nigh critical to the daily usage of an OS. And, I believe, it should have been enabled ages ago (even before the Tracker was enabled) and finished long before now.
In my thinking, you put all your fundamental OS “blocks” together, THEN you refine them and paint them and whatnot (stablize everything). THEN you add your other toys and pleasantries (GUI, Terminal, Deskbar, games, apps, etc.).
However, given that no one is being paid to work on Haiku (except in rare, certain circumstances), it makes sense that personal motivation and interest is all that’s really keeping HaikuOS moving forward at all. And, hence, why people like me complain about certain things we figure should already in place and working, while others (like you) keep chanting the “it’s pre-Alpha” mantra endlessly, trying to excuse the weaknesses that I don’t think should even BE present, all these years later, since Haiku was first conceived.
The more you want something, the more adamant you are about it. The more you talk about it. The more you want others to talk about it. And, if you don’t have the means to get it, the more you complain about that very situation!
I love Haiku and what it represents… but that doesn’t mean I necessarily LIKE how long it’s taking to get to a truly usable state! I’d pitch in, as fervently as Hugo Santos did months ago (where’d he go, anyways? Did he finish his part or burn out or what?), if I knew how to program, but I can’t.
And that frustrates me terribly!
“Fixing fundamental problems after the fact is not easy.”
Amen! I couldn’t agree more!
“The source that is added NOW is routinely not a) SMP clean, b) 64-bit clean and c) reviewed securitywise [like extensive use of unsafe C functions that not even windoze use these days – I’ve been told BeOS was very good at this part].”
While I don’t know the specifics of the above, I think an approach like this could be taken:
1) HaikuOS looks at the hardware.
2) Anything it doesn’t recognize, it states this and says what it is (in a log).
3) It drops down to the most basic functionality necessary, to boot.
64-bit Ubuntu Linux (Feisty Fawn) boots perfectly on my AMD Athlon64 X2 4800+ (Socket 939). Why can it boot on my system and Haiku can’t?
I think the problem with Haiku is that it sees foreign hardware and just stops. To my knowledge, the Athlon64 series WILL work in 32-bit mode. And Haiku works perfectly on my Athlon XP 2000+ rig.
If the North/South chipset is the issue, then drop down as low as you have to, to boot on the thing. If I’m not mistaken, you can boot MS-DOS 6.22 on ANY system. It just works. Why? Because, DOS works on such a primitive level, the higher-end functionality of a given chipset isn’t even an issue! Sure, all you have is a CLI, but you can probably copy files!
I would be happier if Haiku at least booted on my A64 X2 system and crashed for every possibly reason, than just getting stuck at the Haiku Splash Screen like I do now.
I’d be happier if I was stuck in 1-bit (black and white) 640×480@60Hz resolution operation on the Haiku desktop, than NOTHING.
And, I’d be perfectly happy if I could properly copy files, yet everything else was as unstable as a straw hut in a hurricane! Seriously!
I’d rather have 100% proper function, in a 50% stable state, than 50% proper function in a 100% stable state. I want to be able to USE Haiku, no matter how unstable it may be.
Maybe what is needed is a vote of what the Haiku Community wants from the developers. Unless, of course, we’re less important than what the developers want to do… which could be the case as well…
Eh, whatever…
Now, I have also mentioned on #haiku that I was able to reproduce the hangs in VMWare with a 2-cpu VM on a AMD dual core. How hard can *that* be to reproduce for a kernel developer?
I haven’t seen that here in VMWare Player on my windows box running on an AMD64 X2 5600+… is there something special you must do to enable multi-cpu support in vmware? I only see a single CPU listed in Haiku.
Does one need to use vmware server for this?
The rest of your comments aren’t worth responding to at this point. (and Thom already effectively did)
Edited 2007-07-31 15:29
“I haven’t seen that here in VMWare Player on my windows box running on an AMD64 X2 5600+… is there something special you must do to enable multi-cpu support in vmware? I only see a single CPU listed in Haiku.”
Yes, it’s an option on the VM (I used VMWare Workstation 6)
“The rest of your comments aren’t worth responding to at this point.”
Ignorance is bliss, isn’t it.
#haiku today:
[10:24:23] <umccullough> all these really stupid comments on OSAlert are just helping me feed my comment score
o…k…. I leave now.
I guess you’re not familiar with concepts such as humor and irony.
“The only reason you list SMP as a fundamental problem is because your machines are all SMP (I don’t classify dual-core as SMP, but that’s just me, I’m weird). SMP is, by no means, a critical feature in a pre-alpha state”
Well, with that kind of claim, I have to comment.
And, yeah, you’re weird For all intents and purposes, multicores are SMP’s.
There is a couple of projects out there called Linux and FreeBSD. They initially did not have SMP support and retrofitting it was an obsolute pain. FreeBSD is still struggling. It is YOU Thom that obviously have a lack of understanding about OS development. The majority of machines beeing sold now are SMP (or multicore if you wanna split hairs), and not factoring that in from day one in an OS project these days is just a sign of incompetence.
Isn’t this kinda obvious? I know Newos was designed with SMP in mind and was even 64-bit clean (that’s easier to retrofit, admittedly, especially since all the elf64 relocation stuff is well compartmentalized)
Edited 2007-07-31 16:04
The majority of machines beeing sold now are SMP (or multicore if you wanna split hairs), and not factoring that in from day one in an OS project these days is just a sign of incompetence.
*sigh*
Haiku is a re-implementation of a 7 year old operating system, and it’s in a pre-alpha state. It’s not a legacy-less operating system like SkyOS or Syllable, and hence it has a lot more factors to take into consideration.
On top of that, your claim that Haiku does not have SMP support is… Odd. BeOS is a multithreaded environment, with very good SMP support. The fact that it isn’t working very well at this point is to be expected, especially seeing that BeOS was written in a time where multicore was not available.
“On top of that, your claim that Haiku does not have SMP support is… ”
WHERE THE HELL do I claim Haiku does not have SMP support? Point me to the line, please. I started my complaints about Haiku not BOOTING IN SMP mode, but sort of works with the non-SMP kernel.
With a bit of thinking power, you should then realize I do know Haiku is supposed to support SMP. I’m saying it’s severly broken.
If you are thinking about what I said about retrofitting, I was refering to NEW code (which is clearly not SMP safe in many cases)
In fact, I used a lot of BeOS MT kits in the past, so don’t do go there…
Edited 2007-07-31 17:07
WHERE THE HELL do I claim Haiku does not have SMP support?
“There is a couple of projects out there called Linux and FreeBSD. They initially did not have SMP support and retrofitting it was an obsolute pain. FreeBSD is still struggling. It is YOU Thom that obviously have a lack of understanding about OS development. The majority of machines beeing sold now are SMP (or multicore if you wanna split hairs), and not factoring that in from day one in an OS project these days is just a sign of incompetence.”
“If you are thinking about what I said about retrofitting, I was refering to NEW code (which is clearly not SMP safe in many cases)”
Could you elaborate more on this ? If you really know where the problem are, could you please open some tickets in our bugtracker ( http://bugs.haiku-os.org ) ?
Predictor
Almost all processors sold today are multicore ( 2 core design ) and 4+ cores are here too so SMP is very important ( as you stated also ).
Haiku will *eventually* work with *most* SMP systems but not until it reaches Beta or Release Candidate status. That means not until Haiku is ready for a public release. In its current state it is mainly for people to play with and test/check out. It also does have SMP support built in because it works on Axel’s systems though it needs to be further refined to work on more SMP computers.
If you *really* want to run and use a BeOS system today then you should look at finding a copy of Zeta 1.2 ( or newer ). This *probably* will work with your multicore systems. And let you use BeOS until Haiku is ready. You can also try BeOS MAX too which may work. You shouldn’t expect much from Haiku until R1 ( which seems to be 2 years away from having a release ).
Zeta or BeOS R5 ( or 5.1 ) are further ahead right now than Haiku. When Haiku reaches R1 it should have the same or better performance and stability as BeOS R5.
“…. you *really* want to run and use a BeOS system today then you should look at finding a copy of Zeta 1.2 ( or newer ).”
Thanks for the info, but…
I’m only beeing an asshole to hopefully open the eyes of some Haiku devs which will promptly look into the issues I’ve mentioned, cuz’ I think they are so damned important. Sorry .)
Totally uninterested in zeta, yellowtab and any derivatives thereof for non-technical reasons. I want Haiku to succeed.
Having a hard time responding to the correct post?
I’m not Thom, btw.
[10:24:23] <umccullough> all these really stupid comments on OSAlert are just helping me feed my comment score
– glad you were able to locate that…
That’s not saying it doesn’t support SMP… read the rest of what I said. The original Newos SMP stuff is fine and dandy. It’s what happened with the Haiku fork (essentially a new OS) which is the problem.
Why do you argue against Haiku having SMP issues?
Why do you argue against Haiku having SMP issues?
I am not saying there aren’t any problems. Heck, Haiku’s SMP support is kind of, well, fcuked up at this point. However, as I and others have been saying for 63 comments long, Haiku is in a PRE-ALPHA state. Meaning, it’s an incomplete and unstable mess. You can’t expect an operating system in a PRE-ALPHA state to function well on anything else than the hardware the developers have at their disposal.
Haiku is built with SMP support in mind (just like the original BeOS) but currently, that SMP support is anything but stable. Big deal, it’s pre-alpha software here.
You are turning this into another Luposian bug – a bug that if fixed, magically makes Haiku ready. Ready as in, the broadest sense of the word. This, of course, is utter and utter bogus. SMP shouldn’t be a top priority at this point in development. Haiku has other problems to tackle that are way more important.
Like being able to self-host.
Beeing in pre-alpha is no excuse for going in the wrong direction.
Beeing in pre-alpha is no excuse for going in the wrong direction.
And what is that wrong direction, exactly? Haiku’s main milestone to reach, for now, is self-hosting. And if you know anything about operating system development, you’d know that’s way more important than SMP (in a pre-alpha system).
“You are turning this into another Luposian bug – a bug that if fixed, magically makes Haiku ready. Ready as in, the broadest sense of the word.”
Hey, thanks for mentioning my name! Makes me feel kinda special, amidst all this craziness!
I am asking for a feature that, I believe, should have been enabled and working (correctly) years ago. I do not, for even a nanosecond, believe that fixing this one “issue” (missing functionality or whatever) will make the OS ready for prime time. That’s insane!
BUT!
It would enable *ME* to be able to USE Haiku between revisions and feel I was actually USING Haiku, not just tinkering with it for 5-10 minutes, every few weeks, after downloading the latest revision.
I would rather be able to correctly copy files, even if the OS took a stability hit for it. I would rather be able to do EVERYTHING an OS is supposed to do (basic core functions; file copy, video display, etc.), and have Haiku crashing on me for other reasons, than to have a 100% stable OS that can’t do do what little I want to do with it.
But this is a hoary old argument… to say the least.
“And if you know anything about operating system development,”
I’m not getting across my concerns to you; you don’t seem to comprehend what I’m trying to say. With a bit of luck, actual Haiku developers do.
Self-hosting a broken system isn’t much of goal.
I perfectly well comprehend what you’re trying to say. You’re saying that because Haiku doesn’t boot on your hardware, it’s a piece of crap. It’s a common – but invalid – argument among people trying out different operating systems.
Being able to self-host facilitates development, thus making it easier to fix bugs and other breakage, thus leading to a bigger chance your pet bug gets fixed.
“Being able to self-host facilitates development, thus making it easier to fix bugs and other breakage, thus leading to a bigger chance your pet bug gets fixed.”
This makes a lot of sense *AND* I agree with you!
However, there’s one problem… how do you “self-host” on a system that Haiku won’t boot on (such as my A64 X2 4800+ system)? You can’t SEE the hardware, to fix it, if you can’t get boot on the system, period.
At the very least, in a situation such as mine (on that particular system), Haiku should be able to list the entire bunch of hardware it doesn’t recognize, and make it a text file you can then submit to the devs of Haiku. I think skipping the whole Haiku boot screen and showing the raw text of what Haiku is doing would be beneficial in this case.
Haiku, under NO circumstances should just “freeze up”, when it can’t boot on a system. It should show you what it can, record what it can, and tell you what it’s done… even if you’re just left at a KDL screen or something!
I could much more respect Haiku if the Haiku boot screen disappeared and I was presented with the KDL screen and a stream of text showing me what Haiku has done (or tried to do) and then given a friendly:
“Haiku was unable to boot on this system. It has documented what hardware it did/didn’t recognize and saved it to a text file on the hard drive. Please submit this text file to a developer in hopes that your hardware will be bootable in the near future.”
Or, maybe (if possible), boot in some crazy, lowest-common-denominator (LCD) confioguration that at least gives you a desktop you can use, even if it’s as slow as molasses, in 16 colors, at 640×480 resolution or something.
The arguments you’ve put forth so far don’t come remotely near substantiating the claim that Haiku is “broken.”
“The arguments you’ve put forth so far don’t come remotely near substantiating the claim that Haiku is “broken.”
For me, it’s broken for all practical purposes. But that’s fixable, the same way it was for bryanv since axeld “happened to have the same hardware”. Maybe someday, axeld will have laptops similar to mine…
What may not be fixable is current code growth’n’rot. No security reviews (that’s VERY apparent and nobody seems to care), and according to Thom, thinking about SMP and 64-bit cleanliness isn’t important at this stage, while self hosting is. What a horrible strategy (the first few months of self-hosting invariably means an unstable *development environment*).
“What may not be fixable is current code growth’n’rot. No security reviews (that’s VERY apparent and nobody seems to care), and according to Thom, thinking about SMP and 64-bit cleanliness isn’t important at this stage, while self hosting is. What a horrible strategy (the first few months of self-hosting invariably means an unstable *development environment*).”
I also find it a bit annoying when developers are “stepping on each other’s toes”, code-wise. Something works one day and then, a few revisions down the road, it doesn’t. This has happened many times. I’ve seen comments by some devs in CIA, of “Oops, I broke the build…”.
Hopefully, at a time in the not-too-distant future, a very unique version of Haiku will emerge. Familiar in look, but radically different in approach. But key pieces must fall into place first. Just wait… and watch.
If you don’t think outside the confines of “conventional wisdom”, if you don’t envision things that go beyond what others think is all there is… you can never exceed the designs and paths of the present. But you must believe it can be done, before it can be done.
Just look at where Apple is, today, because of ONE man! His vision and determination and leadership brought to us EVERYTHING that Apple had made, since he returned, to today.
It’s not impossible… you just need the right person.
So in other words, it’s not broken – except in your own highly-specific circumstances. And certainly not any more “broken” than should be reasonably-expected of pre-alpha software, which is – by definition – unstable and incomplete (or “broken,” if you prefer).
It may not be satisfactory to you personally, but I think the reality is that a project with the ambitions of Haiku and existing limited resources needs to take a “learn to stand, then learn to walk, then learn to run” approach.
Why would you expect otherwise?
“You’re saying that because Haiku doesn’t boot on your hardware, it’s a piece of crap.”
You must be reading impaired. Re-read my posts. I said I’ve studied the code in addition to trying to make it boot. I have valid concerns about the actual reasons for the problems i’m experiencing.
Stop treating me like a noob.
Stop acting like one.
Speaking as an impartial 3rd party, I’d advise you to quit now while you’re still alive. Because you’re coming off as a total idiot.
I have valid concerns about the actual reasons for the problems i’m experiencing.
Maybe I missed it the first time around, what was it again? All I saw was that it didn’t work and the devs wouldn’t bother fixing it for you.
Wow, seeing the amount of comments in this thread and all frustration about where Haiku works or rather how it works is amazing. I could never have imagined that an OS which clearly isn’t ready yet could draw such attention =)
My take on all this is that Waltercon was prepared a bit sloppy, let’s have this as a lesson for the future. Perhaps a better way would’ve been to have a 2 months deadline from when everything is arranged and announced til where it’d take place. also preregistration and confirmation before finalizing everything would be wise. Then the “scale” of the event could properly have been made.
Anyway, to the File Copying issue and such. My take on it is that developers who work with Haiku right now are fully entitled to do what they please really. IF you have any specific wishes there is a perfect forum available (again) for making things ready (for you that is). Haikubounties is the name of it, and I’m confident that donations is more than welcome! (Webcore is being funded this way)
Regarding information, priorities and metaphores to building houses. I believe the strategy differs quite a lot from building a house. First of all, having somthing which is “self-hosting” will allow development at a lot higher rate than what is possible today (AFAICT). Secondly, the unsafe functions or lack of focus on flexibility right now is most likely a tradeoff to lack of developers. There is only so much that can be done with so few people. It is obviously supposed to be debates about this (as it is in Linux and BSDs) and that is simply how it is.
Now you have a couple of choices no matter how you are
1. sit by and watch and be happy with the progress being made and have Patience (yeah, as if that is a commodity ware these days).
2. Start coding, code talks and debate walks…
3. Donate money, either to Haiku foundation (preferably) or to Haiku Bounties for any other wishes
4. Walk away and do use whatever you want.
Me personally tend to go for option 1 and sort of something on the side, I try to talk to people about Haiku and show the world there is another option (partially through http://www.teamhaiku.com ). My best wishes to those that code =)
“Going to #haiku and complaining is as effective as telling it your mom. She probably doesn’t know what you are talking about, and can’t help you.”
Well, I happened to watch the #haiku logs yesterday and did some searching and there was several discussion on SMP not working in qemu and vmware, for instance. Should be easy enough to reproduce.
And #haiku is read by core devs on occation, not only moms (though, there are moms who are kernel developers out there
By the way, as many I REFUSE to submit bug reports to a system where you need to register. Why not do as Syllable, for instance, and allow anonymous postings. It’s the same with linux; anyone can post problem reports on lkml.
You just don’t understand the issue. There is no way to reproduce it on any of the developer machines. There is not a single tester who takes the “problem” to be important serious enough to do a proper bug report. Talking is cheap.
Sure it is. But obviously talking about bugs there often doesn’t help to get attention to the proper person.
Haiku is developed by volunteers. It costs our time to process bug reports, and it is important to be able to contact the submitter of a bug report. Thats the reason why you need to register with a valid email address, before you can submit bug reports. No other personal information is required to enter. We do not allow anonymous bug reports, because we need to focus on getting the required information to fix an issue.
Edited 2007-08-01 11:41