Robert Szeleny has given a short update on the status of SkyOS. “For the last weeks and as it looks like quite a few more weeks I’m in contact with a lot of people talking about their ideas to continue SkyOS development. Unfortunately many people completely underestimate the amount of work required to manage a development team for SkyOS, either an open or closed source version, or completely lack vision and or motivation and or seriousness. Anyway, I’m still reading through my inbox hoping to get in contact with people with real motives and knowledge being able to bring SkyOS to the next level.”
So many people underestimate how much work it takes to manage a dev team that isn’t paid and isn’t all in one physical room. The propblem is, people who know how much work it takes usually don’t want to do it because they understand just how much or your time it woll consume. It is or that very reason that I personally fall into the “wow it’s way to much stress” group.
What I might recommend (from personal experience), is try to have multiple groups and not lump all devs into one. Just like the translation groups are not needing to sync up with the developers all the time, so can you divide up the developers.
1: The core API, kernel, and driver people need to be in constant sync and are hopfully few in numbers (this is the group that needs to closet monitoring and management).
2: User space and base OS aplications can be managed in a seperate group, and need to be in comunication but not to the same level as group 1.
3: Then there is the community and managing patches comming in, ports, ideas, bug, etc… (not as daunting in non OSS aplications). This part can be even more loosely managed. etc..
3 main groups with 3 different maintaners might be necessary. The 3 team leaders need to be in constant sync, but it makes it easier to do it in 3 groups. I have done it that way for my major endevours (though non as daunting as a full blown OS), and it has served me well.
Good luck Robert,
-Chris
I hope it continues. Developing using the SkyOS APIs is great. I was able to port a gaming library in 2-3 days (graphics, sound). It’s easy to pick up.
The author has done an incredible OS with limited time, etc.
He needs a mini-team to continue. He could still develop the vision and do special projects, but leave the core development to a group of developers.
Does anyone use it as there “default” desktop(prefered os)?
The problem Mr. Szeleny seems to have is letting go. It’s remarkable that a one man operation can go so far and achieve so much as Mr. Szeleny has, but the times have overtaken the viability of the one man endeavor.
There are basically four options:
1.) Continue development with “third party” developers on the team.
2.) Sell SkyOS lock, stock and barrel to a party that might or might not develop it further.
3.) Open source it and leave the (bulk of the) development to a SkyOS community.
4.) Shelve it and let SkyOS join the other OSes who have fallen by the wayside.
Only option four will leave Szeleny in full control, albeit with a Pyrrhic character. All the others will mean letting go of complete control. Only option one and three leave (some) room for Szeleny to have a say in what is and is not the vision behind SkyOS.
To let SkyOS live, Szeleny will have to step down as the only brain behind it.
One-person-shows are though to sustain over a long time unless they are joined by a large crowd of paid or volunteer “slaves”.
-> Paid slaves come from the initiator having gone into the market place and raised a whole bunch of money. As a reminder, Apple, Dell, Microsoft, and quite a few others have been essentially one-person-shows for most of their histories. The relationship is based on power hooked through one’s pay cheque.
-> Volunteer slaves come from enthusiasts willing to work hard because they share a similar vision to that of the initiator. The relationship is symbiotic and the volunteers are free to join/leave the project as they please.
In both cases, clashes of opinions (and at times just personalities) cause a group of slaves to depart and form their own project. Going through the history of the various Linux distributions and how many of them came into existence can be quite hilarious in this regard.
I have followed the SkyOS development since introduced to it in an article on OSAlert. Paying for the privilege to contribute to the beta development program was something I was not willing to do. However, many good concepts (and the code bringing them to life) make SkyOS what it is today.
Mr. Szeleney is facing a fork in the road with respect to SkyOS:
i) Sharing ownership and the direction of the project with those willing to contribute code to it. This is something that is particularly difficult to do for many technical visionary individuals.
ii) Merging the project with another one with some similarity of the foundation concepts. For example, the SkOS File System traces its seed to the Be File System. How divergent has it become from the original seed may determine how its code base could be merged with other projects based on a similar file system concept (Cosmeo/Syllable being forked from Atheos? Haiku being the open-source rebirth of BeOS?)
iii) Abandoning the project altogether – whether with a full general release of the source code or allowing it to rot.
Only him, and a few close long-time friends/partners, can take one of these forks. Hopefully, it will be one which will benefit projects based on similar concepts.
Well, I can only say one thing about this… Robert is a person that can set a very high and unreachable standard and that will be the death of SkyOS.
He kept it closed sourced, so he could make sure it would be developed according to his standards.
He sold beta’s to fund the project with a prospect of making a perfect product.
After 8 beta’s (he made his mind up he was nearly there if I recall it right), he started adding new features and more and more. Improving and improving solid parts of the OS, instead of simply releasing it and making sure it would get the community to evolve further.
Now he wanted to build in virtualisation and decides it’s to much…
You can say what you want… But now he’s even making a fuss about who should develop further. I think it will take about a year to find them… At least, he probably set that timespan… but like the release, everyone just have to wait and wait and wait…