Microsoft is at a crossroads, and the operating system that follows Vista will likely mark a serious break from the past. That is, if Microsoft can figure out how to do things differently. Also, the ‘MEDX’ team shares desks, ideas on the look Microsoft wants.
The Next Windows After Vista; New Model for Device Team
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSAlert.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
The system has been growing with every version for the past 20 years. It’s become so bloated with old code and features that the drawbacks (security holes, resource consumption, regulatory ire) outweigh the benefits.
…they hired Captain Obvious.
Indeed. Adds nothing new.
True, I find especially funny the ‘if Microsoft doesn’t improve things, it is dead’ tone of the article: Microsoft currently earns a tax on every PC sold, so even if it did nothing excepts patching its current software it would still be a very, very rich company.
Sure there can be a few people who migrate from Windows to other OS because of some features of the other OS but currently, apart from security factors, there isn’t any real reason to switch to other OS.
And for compatibility and games, there are many reason to keep using Windows..
In the technology business, a company is in big trouble if it doesn’t keep on top of things. IBM has never been anything but a huge company (well, not for a LONG time anyway), but it got in big trouble towards the end of the eighties/beginning of the nineties.
As for “no reasons to switch,” well I suppose we’ll just have to agree to disagree!
Edited 2006-09-06 11:09
Nice words, but considering how much of Vista has been chopped at some point, please excuse me for being highly skeptical.
There’s really no reason Microsoft couldn’t have done this back in 2002 when Windows XP was released. Only now are they admitting what everybody has known for years. It’s also partly an excuse, screaming “backwards comptability.” Look at what WINE has been able to do. Then imagine if Microsoft’s knowledge and resources was applied to doing the same thing. Apple did something similar with their OS9 compatability environment.
It’s not a technical problem. It’s a social problem.
Ah, but! WINE provides compatibility for a small number of applications, and even in those there are often bugs.
OS9 provides full compatibility at a speed cost, and a fairly sizeable disk space cost (not to mention the cost that a customer must buy both systems if they don’t already own them).
Windows compatibility is so ridiculously high that when anything at all fails to be compatible people think it’s a big deal.
I’m not saying it’s a good thing. I’m just saying the level of compatibility is probably what really gets them.
They don’t say “let’s not make ISV’s rewrite” they say “let’s not make ISV’s recompile or repackage or even modify in any way.”
I can say with the utmost certainity based on years of prior experience that neither WINE or Classic are bullet proof. While both are very useful and functional, I certainly wouldn’t depend on them for anything critical.
Maybe Microsoft should consider following Apple’s lead and take BSD and graft on a proprietary GUI and multimedia framework like Direct X. It sounds like they want to move to a modular design where libaries of code are shared between applications. Their large monolithic design has finally caught up with them. Hmm. Sounds a lot like the advantages of using linux.
I wanted to write something very similat to what you say, as first post. But then I changed my mind, here it could have sounded like flamebait.
And indeed, Apple did that and now it doesn’t look as if they need to change anything radically any time soon.
Wow, what a great demonstration of lack of Windows understanding you’ve displayed! I’ll be about as concise as possible:
1. Windows has oodles of modules that are reused: that’s what DLLs (Dynamic Link Libraries) are, which are a different name for what a lot of other systems would call a .so (Shared Object).
2. Direct X is based on COM, which has its entire reason for existence (COM) as being what it is an acronym for: Component Object Module. COM is merely a more precisely defined version of DLL that also incorporates a way to include multiple versions of binary interfaces that can change over time. In other words, a single COM object (a DLL, regardless of the extension put on it, but is often .ocx for visual stuff, often .dll, but can *also* be grabbed from a .exe, even one that is able to boot independently!) can have the same name for an interface to code it supports, but you can explicitly ask for more than one revision of that interface, which may also have different semantic properties. In the ideal case, COM was supposed to help solve the “DLL Hell” problem that can also happen with other platforms, if applications truly did share their shared objects with every other application that could use them.
3. A lot of the more recent GUI controls for Windows are based on COM, which can make things a bit odd, because how you have to use them has its weird differences to worry about. However, it’s…quite modular.
5. The big problem with Windows and why some people perceive it as being monolithic, when it really isn’t any more monolithic than say, Linux combined with any windowing environment, is that the standard guaranteed-to-be-there APIs that exist are *HUGE* and very powerful, but have evolved over many major revisions of the OS, with Microsoft often going to extremes to maintain backwards compatibility, including maintaining behavior in more recent versions that is identical to what’s clearly buggy behavior in the original versions, precisely because applications depend on the buggy operation. If Microsoft fixed the implementation of everything (because there’s so much available functionality, and by the law of averages, so many odd quirks) to match what you’d expect it should have been originally, it’d break a lot of applications.
Mac OS X doesn’t have all this old baggage, but it also doesn’t have as much built-in functionality, either, but has some interesting things in it that Windows doesn’t have, instead. Windows XP and Win2K3 Server still support Object Linking and Embedding (OLE) and DDE, while OS X simply doesn’t have that. There are a lot of other things, too, that fall under that heading, that Windows mutations support that the other OS’s don’t have, and vice-verssa, too, that makes the platforms unique. Windows is very powerful because of how modular it is, combined with backwards compatibility, while also being restrained by the backwards compatibility, combined with the confusion that comes from having many cooks making the soup over many years, often without a common over-reaching theme throughout time for how things are supposed to be done. Is this a good or a bad thing? A bit of both, really: if Microsoft never tried anything different, people would bitch that they don’t try to advance things, and by trying something different, they also make things more confusing over time. <sigh> there’s simply no way to win and have backwards compatibility, and if they break backwards compatibility, that’s not always great, either, because of how much software exists that you never see on a shelf that businesses depend on. I can’t help but think they’re making VPC available for free as their way of providing an “XP Classic” environment, because they’ve bitten the bullet hard enough with Vista that a lot of software won’t run on it, and they know they need to make it possible for customers with older software to work, while still giving them a reason to consider getting Vista.
You are brillant Mr. Thompson. It’s rare to see someone who really knows Windows these days.
Anyway, all old school Windows programmers will tell you that Windows is the best OS because it is.
You are brillant Mr. Thompson. It’s rare to see someone who really knows Windows these days.
Could that be because Windows *actively discourages* “really getting to know it”?
Anyway, all old school Windows programmers will tell you that Windows is the best OS because it is.
What brilliant reasoning.
I love how you start of saying the response will be concise and what follows reads like a lengthy MS technical document. I would love to see a more verbose version of your conciseness but I am afraid I would be reading it until winter time.
Your points are well made but it reads like a lengthy excuse. A friend of mine develops code for the MS Office Suite and according to him the spaghetti code which communicates between the Office modules and the OS is incredibly mind numbing to work on since the complexity is so high. The Windows OS culminating in the Vista release is the height of this complexity where it is showing to be almost unmangeable – hence the continuously pushed back schedule. Perhaps they have tried *too* hard to maintain backward compatibility at the risk of destablizing the OS.
love how you start of saying the response will be concise and what follows reads like a lengthy MS technical document. I would love to see a more verbose version of your conciseness but I am afraid I would be reading it until winter time.
The quick and dirty answer is that Windows is modular, thats what all the DLL and OCX files in your system32 folder are about, they are modules.
Now that dosen’t mean that the *management* of those modules is clean or straight forward. COM (component object model) requires that a software interface be defined for an object, this is where a big part of the registry nightmare comes into play, these interface descriptions get screwed up and windows start acting craazy.
A friend of mine develops code for the MS Office Suite and according to him the spaghetti code which communicates between the Office modules and the OS is incredibly mind numbing to work on since the complexity is so high. The Windows OS culminating in the Vista release is the height of this complexity where it is showing to be almost unmangeable – hence the continuously pushed back schedule. Perhaps they have tried *too* hard to maintain backward compatibility at the risk of destablizing the OS.
I’ve got a friend at MS also and I’ll have to agree with you. The OS has become complex, as a motherf*cker for sure.
Backwards compatability is a big part of the problem as you’ve stated. I was personally hoping for something along the lines of MS using a win32 ‘classic’ layer driven by Virtual PC and a clean API in vista based on the .net framework but I’ll have to hope its in the next release of windows.
I was personally hoping for something along the lines of MS using a win32 ‘classic’ layer driven by Virtual PC and a clean API in vista based on the .net framework but I’ll have to hope its in the next release of windows.
Which simply doesn’t materialize in time. When it does, it is already old news for other systems. MS is a follower, not a leader. You get good quality stuff from last year.
standard guaranteed-to-be-there APIs that exist are *HUGE* and very powerful,
The API’s in Windows aren’t all that powerful relative to their size. They’re just huge, full of features that almost nobody needs, and could better be addressed using compositions of orthogonal APIs. If the APIs were better designed, they’d be a fraction of the size.
This is a structural problem at Microsoft. They have so many programmers, the API designers stuff in anything that comes to mind, instead of judiciously including minimum subset necessary to accomplish the goals of the API.
APIs like OpenGL or POSIX or Berkeley sockets would never have come out of Microsoft. “Represent a display list as an integer? Absurd! It should be a full COM object! In fact, everything should be a full COM object! And while we’re at it, let’s include support in every DX DLL for every other DX DLL ever released!”
Edited 2006-09-05 02:03
heh, sounds like a software version of the cisc vs risc debate
Vista: 50,000,000 LOC
Debian 3.1: 213,000,000
Which is the bloated OS?
http://en.wikipedia.org/wiki/Source_lines_of_code#Measuring_SLOC
The difference is *flexibility*
Who on earth will install Debian with every application?
On the other hand try to install Windows with a 100 MB iso, as you can with a Debian netinstaller (and yes, you get a fully working OS)
I do realise that Linux-hating is your religion, so despite my willingness to debunk “religions” such as Scientology I’ll try to be gentle.
Those figures which you quoted, if reliable, no doubt refer to (a) the size of Vista, when installed, on the one hand; (b) the total size of all (is it 18,000 now?) packages included in Debian, on the other. So, which looks bloated now?
———-Vista: 50,000,000 LOC
Debian 3.1: 213,000,000
Which is the bloated OS?———-
Wrong.
Add up the lines of code for Vista, plus MS Office, plus Nero, plus Exchange, plus MS-SQL, plus VirtualPC, plus MS Visual Basic, plus throw in a handful of games.
Now tell me which is more bloated.
Debian like most other linux distros come with alot more than just an OS with basic/cut down features.
Edited 2006-09-04 23:07
Since when does lines of code == bloat?
since ;
when does ;
the physical number ;
of;
lines ;
provide any ;
indication of ;
compiled code ;
effectiveness or ;
efficiency ?
Of course, in the good old days you got paid by the line…
Please, please, please, please; if you’re going to bash GNU/Linux, please do it right.
If you want to compare it to Debian then it would be:
Windows Vista Ultimate + All of the opensource that Debian has, running under services for UNIX.
Lines of code isn’t the problem; its old crusty crap that should have been purged faster than a bolemics last meal.
…I mean, look at UNIX (in a general sense). But I would guess that no modern UNIX includes support for DECtapes, PDP-11s, or 1.5MB drums. And it also adapts well: I doubt most use of tar relates to Tape Archives these days. Windows, otoh, is still stuck with drive letters. ugh.
Drive letters would be the easiest solution. Drive numbers would be ugly, drive names too long to type and a single file system tree confusing for most users.
Users are not as stupid as you make them to be.
if the CD drive is G: or /cdrom makes absolutely no difference to a user. Especially if the Explorer displays a nice CD icon instead of a folder icon at the location.
I think the historical mountpoint /mnt/cdrom in many Linux distros is not that easily findable. But even my mother managed to memorize that.
If MS wanted to skip their drive letters, they could be easily doing it, and users would just say: “See there, Windows has no drive letter, but it works similar as before.”
How would “drive numbers be ugly”? AmigaOS and RISC OS use(d) them just fine. And they certainly provide more scope for expansion than the 26 letters of the alphabet.
Anyway, who types even drive letters on Windows these days?
Anyway, who types even drive letters on Windows these days?
I sometimes do, when I’m feeling self-sadistic and go to “work” in MSes DOS-emulator.
[QUOTE]Windows, otoh, is still stuck with drive letters. ugh.[/QUOTE]
Not really. Since NTFS3 (2000) you’ve been able to mount partitions/drives pretty much anywhere you want in the filesystem. I have my second partition mounted at C:TEMP, for example.
Having drive letters for removable media works pretty well. OTOH, it’s a PITA accessing removable devices from a console/terminal in *nix.
bill gates ran out of ideas. now ray ozzie will try to incorporate his, does not matter what new innovations come about, because UNIX will remain the one true OS.
Edited 2006-09-04 23:34
The trick to Webifying Windows will be keeping essential components on a PC’s hard drive, while moving some user files and the software that interacts with them onto servers in Microsoft’s data centers.
Now there’s a great idea – add some “trusted computing” and you’ve got a guaranteed winner (I can even imagine Linus switching to Windows when this is released).
Makes me wonder how much Microsoft would charge us to use the datacenters – I’m sure it’d be a reasonable price, and that the price would decrease rapidly once Microsoft have the power to decide who does and doesn’t have access their own files…
MS should just ditch it and put all the (wasted) millions they spend on Windows into Singularity.
http://research.microsoft.com/os/singularity/
http://en.wikipedia.org/wiki/Singularity_(operating_system)
(That parentheses is part of the link, don’t know why it gets cut out)
Edited 2006-09-05 00:24
If people see Microsoft moving to Singularity because Windows is a PITA, that might just encourage them to move to Linux (and by “them,” I mean corporations and ISVs). After all, if Microsoft don’t have any confidence in their flagshit^Hp product, why should we have any confidence in Microsoft?
EDIT: OTOH, Singularity might just end up being a black hole we all fall into.
Sorry, couldn’t resist
Edited 2006-09-05 02:21
“After all, if Microsoft don’t have any confidence in their flagshit^Hp product, why should we have any confidence in Microsoft?”
True. Singularity won’t ever exist as a true OS. Nobody would use it. There’s Linux and works fine. If Microsoft ditches Windows because it cannot support the mess anymore, then it is done forever. Nobody will ever want something from Microsoft, companies all across the globe will be furious and start collapsing while those that know well will switch to open source. The others would stay using Windows for years until they finally find a way to migrate to open source, possibly loosing money.
But it doesn’t take a lot to bring Microsoft down. If Windows Vista is a shit nobody would really want (I hope so), then there will be some pressure on Microsoft’s stock holders, employees will find (as they do now) that their company is no longer a good job or has a good future, will go away, ISVs and developers will start looking at Linux a bit, just in case. The world will choose what it thinks is more viable for the future, and those that don’t get in the train will suffer the consequences.
I hope to be alive when this happens.
Well, Microsoft could turn around and start doing things right, like IBM was forced to do.
They could become the leading Linux distro. Of course they would need to be at least as open as Linspire.
Then they no longer would have to code their own OS kernel stuff, but could still ship their proprietary addons like media player and internet explorer.
Now that I think of it, I wonder why they did not do it already (of course: backwards compatibility).
They could become the leading Linux distro. Of course they would need to be at least as open as Linspire.
Then they no longer would have to code their own OS kernel stuff, but could still ship their proprietary addons like media player and internet explorer.
Microsoft Linux??? sounds scary!!!!!!!
Edited 2006-09-05 11:20
Microsoft Linux? Doubt it’ll ever happen. Even with proprietary icing, like OSX, it would be a bigger vindication of Linux than Microsoft bankruptcy. Microsoft have to go bye-bye someday, whether Linux is out there or not.
Microsoft BSD is more likely, unless the people who have left/are leaving now are the only ones who hate the GPL.
Apple, as he said did move from an OS to another OS but its older OS was completely a different OS that OSX could be considered an OS from scratch rather than a complete rewrite.
Now, what he said is that they wish to make a rewrite of windows without considering compatibility, but he never mentioned to rewrite or redesign the OS or to even depend on another projects which is more stable and free to use like GNOME, GRUB and others in their next windows (if that would ever be called windows).
Last time I tested vista (build 5552 a pre-RC1) after installing Alcohol 120% the OS didn’t complain about installing it, it didn’t stop me from downloading it, then when installation was complete, BSOD appeared, immediately I understand it is the software I have installed was not compatible, then I booted to safe mode and another BSOD then I booted in safe mode with networking then again a BSOD, so how could a person less than a technician uninstall the application or the service of that application when windows doesn’t have a decent real “Safe Mode” apparatus to allow administrators to troubleshoot them?!
Windows suffer not only from what Mr Ozzie just said, It also suffers from being dumb and designed by dumb people to allow even your kid to mess it up to a point of no return but a complete reinstall.
Design-wise windows suffer horribly from GUI to API bad designs and if they are not gonna change it and start to protect their OS services and create decent troubleshooter then their OS will fade soon to Apple and Linux emerging technologies maybe not now but in the very near future.
All I saw in windows vista ( in comparision with server 2003), is the good looking icons and some decorations with some fast installation procedures but function wise it remains a crashable OS with more demands for hardware, leaving us no justification to say windows is more cheaper than Mac.
I read subscription based model.
Good thing we haven’t heard that before.
My take is:
have Vista be the last of the old line; have it have a much longer support timeframe.
then, have a clear guideline published with tools telling programmers and ISVs what the new model will look like. just break the old stuff out and go to more cleaner designs.
this way once you have newOS out it can run parrell for a few years and then some. then ISVs can port there apps to it and if everything is dotNET based should make things not so painful.
-Nex6
Linux blows and it will never get up to the standards set by MacOSX, Vista client or Vista Server.
You have a choice, Mac or Windows.
Linux is a choice that some people on here want to force people with but they never can because it is crap.
Linux is a choice that some people on here want to force people with but they never can because it is crap.
I’m sure you’d love to think that, it would certainly make you sleep more easily. But the fact is, if Linux were as crap as you say, you’d just ignore it and let as many people try it as possible so they could see just how crap it was.
Whereas in reality it’s an excellent and flexible OS that offers something the other two don’t, which is why a lot of people use it. People who are already well aware of Windows.
“Linux is a choice that some people on here want to force people with but they never can because it is crap.”
You can not claim it as crab.
I have fedora for a year now and it never crashed with any video codecs I have installed to, Its so stable if the crab and non crab hardware and software manufacturers decided to support.
Look around you -if you can- and you will see IBM, HP and DELL started years ago to sell workstations with linux preinstalled and why? is it because they want to waste their times?, or is it because linux is crab? Neither of course.
I know that linux is not to be mentioned till now when it comes to home/SOHO desktops but for other things it rocks (server/Workstation).
I pray for your eyes to see the light one day!!!
I know that linux is not to be mentioned till now when it comes to home/SOHO desktops but for other things it rocks (server/Workstation).
I pray for your eyes to see the light one day!!!
Please, don’t delude yourself into thinking its ready for mainstream adoption; until they get the kinks sorted out, the hardware support up and more softwrae vendors on board, it’ll remain a niche until that day.
I ask, with that assessment, whats wrong with being a niche? concerntrate on what you’re good at; as soon as your marketshare starts expanding into the unwashed masses, you bring upon yourself a whole wheel barrow of pain and missery for all concerned.
I’d love to see, for instance, PC-BSD, which I think probably, if it has some commercial backing, a much more viable chance of getting to the desktop due to the non-political nature of the FreeBSd community, and the stable driver API which they provide.
For me, I’ll be purchasing a Dell laptop/desktop soon, and putting on it, most likely, FreeBSD 6.1 along with Xorg and KDE; for me, I’d rather the concerntration be on providing the best operating system rather than one that panders to all the problems and issues that end users face, and as a result, dumbing down the whole ecosysteem as a result – lord knows I don’t want to see idiots spread their ignorance into forums wanting to know how they can put a picture of their pig ugly kid for the background, or some user whining because he can’t change the theme to ‘the colours of the football’ that he supports <shudder>
I’d love to see, for instance, PC-BSD, which I think probably, if it has some commercial backing, a much more viable chance of getting to the desktop due to the non-political nature of the FreeBSd community, and the stable driver API which they provide.
Non-political? Hahah. The FreeBSD community is *very* political.
Anytime anyone accuses someone of being “non-political”, you can bet it’s because the accuser and the accused are *on the same political page.*
Non-political? Hahah. The FreeBSD community is *very* political.
GPL, Linux and binary drivers anyone?
Nice to see you avoided the issue of a stable driver interface – but hey, not all of us are perfect
Non-political? Hahah. The FreeBSD community is *very* political.
GPL, Linux and binary drivers anyone?
Allowing people to do whatever with your code is a recipe for forking. That’s why even companies that develop a closed-source model but allow customers to see source keep a tight lid on it. The dichotomy isn’t between closed source and BSD-licenses, on the one hand, and GPL on the other, but between closed source and GPL on the one hand and BSD on the other.
As you’ll know if you’ve ever read any of the BSD mailing lists, the OpenBSD folks might be the most vocal of the BSD’ers who hate binary blobs, but FreeBSD has plenty to go around. Try replacing all your FreeBSD drivers with closed-source ones, and you will see that vaunted uptime disappear down a big, binary hole.
Nice to see you avoided the issue of a stable driver interface – but hey, not all of us are perfect
Nice to see you avoided the issue of hardware and vendor support.
I like FreeBSD, but its most rabid proponents would have you believe that it’s more business friendly than Linux. Given that Linux is spreading like wildfire amongst certain types of businesses (with indications that it is also, slowly, broadening its appeal), if it were more business-friendly than Linux you would expect FreeBSD to be bringing Microsoft to its knees. And don’t say “It’s because of the AT&T vs UCB lawsuit” – I bet half the companies (at least) using BSD these days don’t even know about that, and as we know, plenty are also rabidly opposed to the GPL. Compared to Linux, FreeBSD and even PCBSD adoption is just not happening.
The lawsuite is ancient history, what has slowed down the adoption is the lack of commercial support by way of products and services oriented around FreeBSD – the last vendor, which provided BSD/OS have scaled downt their committment, and coupled with the lack of software and hardware vendor support, its going to be hard.
Linux also has the benefit of hype behind it; if IBM chose to support FreeBSD over Linux; we’d be talking about FreeBSD on the desktop and/or other issues at play – its not about technical superiority in regards to Linux, but being in the right place at the right time, with the right people in the right locations to push Linux to those in companies who make decisions.
With that being said, however, it isn’t the panacea to all of lifes problems; OpenSolaris has helped Sun claw back marketshare, Microsoft keeps pushing forward with Windows 2003 Server; and recent information shows that the growth of machines with Linux pre-installed has slowed – don’t make the assumption that just because its Linux, it can’t defy the most basic rule of business; he who provides the best product, wins more customers, and with the major improvements in Windows 2003 and the rebirth of Solaris, its going to be alot harder in future.
Whatever the reasons for the vastly greater adoption of Linux over FreeBSD, you’ve conceded the point.
And I’m surprised you take that study credibly: It leaves a lot of unanswered questions. Such as “are the people who buy Linux servers buying less Linux servers, or less servers in general?” “Are the people with Linux servers replacing them with Windows?”
BTW, that’s the first time in a long time I’ve heard Windows described as “the best product” with a straight face.
…which didn’t make my point well. The point is, first you say that FreeBSD is better than Linux, but Linux is more widely used amongst business because it gets all the hype, but then you say that the fact that Windows is used more is due to its being better, and that’s why it gets all the hype!
Both can’t be true!
he who provides the best product, wins more customers, and with the major improvements in Windows 2003 and the rebirth of Solaris, its going to be alot harder in future.
Obviously you have some reading comprehension issues; I clearly pointed out TWO good products, Solaris and Windows 2003.
Damn I hate 15 year old linux fanboys who can’t be buggered reading, comprehending, thinking and then coming up with a decent reply without needing to waste valuable bandwidth with needless multiple posts as replies.
Windows 2003 has been out for nearly four years. I can’t see how a four-year old operating system is going to suddenly start making waves.
As for Solaris, even as a Linux user you could quite happily go through life not even knowing it exists, I imagine.