Longhorn Developers @ MSDN 454
ePIsOdEOnline writes "The official Microsoft Longhorn Developers website went live. Content is filled with information fresh from the PDC, and the host of secrecy swarming Microsoft and its next generation Operating System,
Longhorn"
Guys this is a total Win98SE (Score:1, Interesting)
WinFS was supposed to replace NTFS. Now it's just gonna be IMDB (in-memory database) which was stripped from Win2000 at the last minute file-system filter driver, which will be used by Explorer. Neat, but isn't gonna change your business.
The shell theme will change.
The most interesting development, Avalon, will attempt to replace the Win32 api with managed code, *not* Windows.Forms but a new client API. What'll end up happening by RTM is Microsoft will have written a great big propriatary app that you won't be able to use.
There will be minor kernel perf tweaks, but if you still launch > 200 processes you still see random failures from CreateProcess. (Never tried it, have you?)
How are they gonna manage this? (Score:4, Interesting)
I say things up there about 'migration' and 'preparing' and 'interoperability' but I didn't see a way for them to maintain support. Linux can maintain an active beta because people can actually work on it, so they can more easily test it and benefit immediately from that testing.
Microsoft, I've seen many claim, is drumming up support and mostly trying a publicity stunt. The question is, how do you run a 2-3 year publicity stunt?
Maybe they should ask SCO.
Re:Loooooonghorn (Score:5, Interesting)
It's a so-called 'guided tour of Longhorn', which consists of no actual imagery, but rather a gigantic step-list of things for you to click on in your Longhorn alpha, to make you go 'ooooh'.
Just brutal. I mean, if its really a 'bet-the-company' strategy, you'd think they'd splash out just a little cash for a Flash or non-ass-looking PPT prez... or even screenshots.... something other than this. Just looks really amateur.
Um... which of these options do you prefer:
1. Microsoft makes fake demos of things that don't exist yet.
2. Microsoft has the guts not to show what they don't have.
This is a developer site, not a marketing splash. Developers want information, not pretty presentions.
The fact that they are willing to give out information on a product which is years from being finished shows both courage and strategic integrity, if you ask me.
Interesting comment from Bill (Score:5, Interesting)
Windows95 originally was just going to be Windows 4.0--an updated version of Win3.1 Turning it into more than a GUI for DOS, adding multitasking, recreating the GUI, and so forth, was a HUGE undertaking which lead to endless delays. (Win4.0 was supposed to be out in '93; Win95 barely made it into it's named year.) But what threat caused the massive effort? OS/2. OS/2 2.1, the PPC chip, and the Pentium FP math bug got MS good and scared, and they came up with a (relative) miracle.
Now they're saying that they're putting that effort in again. What, pray tell, is the threat to MS this time, hmmmm?
New OS, same old company behind it... (Score:2, Interesting)
Or is there going to be the convenient clause in the EULA which states, "the consumer will be obligated to periodically, by an automated process download and install patches or warranty is void" This could be the OS that finally gets everyone onto broadband/DSL/etc. due to the shear volume that each will have to download. Yay!
Then again, many will try to use this operating system on stand-alone systems, which will probably be some violation of terms of the EULA, where Microsoft needs to know everything you have on your computer and what you're doing with it.
I could make an 'all your base' reference, but as Breathed and others have noted, you can't compete with reality anymore.
WVG? (Score:2, Interesting)
Re:who cares? (Score:5, Interesting)
You know, that's exactly how I feel.
I still run 2k on my home system (mostly out of lazyness really, theres nothing windows-specific that I really rely on, save a couple games on occasion). I still run 2k on all the workstations at work (except mine, which is RH). My PDC is NT4 on a 500mhz box, as it has been for 3 years, and once I get time it will become Samba3 (it also runs Apache and BIND).
I do have a 2k server, which is running our accounting system (unfortunately, the low-cost (under $1k) linux-based stuff didn't meet our needs, linux mid-market ($5-20k) doesn't exist, and the rest is $90k+). We ended up going with a mid-market windows-only solution, but that system is ready to be a terminal server.
I have all the pieces in place, and my ultimate plan here is to switch all our desktops over, once I find something that isn't going to reduce the 'feature set' of our desktops. An upgrade isn't really viable if I have to tell everyone "oh sorry, you can't select printing options when you print anymore, they can only be set in the driver options - which you can't access".
I'm just totally giving up on windows. It's just not worth the hassle anymore. I can't do any sort of automatic app installation, which is one of the things that bothers me a lot. Our office is small - 8 workstations - but it's big enough that it takes a lot of time to go around doing windowsupdate, installing version x.y+1 of whatever, etc. I never found a nice solution that didn't cost a lot of money. (And yes, I know 2k can do it. I've used it at another company, and we had to turn it off because it made things more difficult). This is the sort of thing that I can use rsync and a couple shell scripts for, and have a working solution in half an hour. Flexibility is key: My job is not to be a sysadmin (we're not even a computer-related business), so the less time I spend sysadmin-ing, the better.
Anyway, that kind of turned into a rant, and i'm not looking to fight with any of you MSCE's that are going to try and counter everything I've just said - I've heard it a million times. I'm just trying to point out that if you're going to wait, wait for something worth waiting for. I personally don't see longhorn adding anything that justifies the expense.
Microsoft's Obvious Strategy (Score:4, Interesting)
Knowing that, Microsoft is deliberately drumming up the hype now with an outrageously early beta, leaked screenshots, and surreptitious press releases and leaks about their upcoming features. Why? To get the current installed base excited about the next release, and to quiet any concerns they have that might make them switch in the interim. If they saw no compelling reason to stick around until 2006 they may migrate to other platforms. The leaks and beta try to give them that reason.
Re:How are they gonna manage this? (Score:1, Interesting)
Apparently you haven't seen much of the Windows development community. There really are people who write software that uses the latest Windows APIs and need to develop for them before the OS gets released. There were several active WDM mailing lists going on for years before WDM actually showed up in a release OS.
Linux can maintain an active beta because people can actually work on it, so they can more easily test it and benefit immediately from that testing.
This is a beta test for application/driver developers, not OS developers. If you are developing for a new OS, it's handy to have a copy of that OS before it gets released so you can make sure your software is compatible. It's more about getting the latest stuff out there to test with your applications than it is for you to find every Windows hole and get it fixed before release.
Re: How are they gonna manage this? (Score:3, Interesting)
> No, seriously. How are they gonna maintain an active interest during the next two years of development?
Two years? In general you should double the estimated time-to-release for IT products.
And I thought they were saying 2006, so the problem is how to maintain interest for 5-6 years.
> Microsoft, I've seen many claim, is drumming up support and mostly trying a publicity stunt. The question is, how do you run a 2-3 year publicity stunt?
When you have several billion dollars to throw at it, and the press hangs on your every word to begin with, it shouldn't be that hard a problem.
Hell, how long has it been going on already? When was the first time you heard of MS Longhorn?
Re:Wise choice (Score:5, Interesting)
There is other new tech going in as well.
This really is a big step, and Microsoft is making it public right now so developers can get on board early in the game and make suggestions or comments on it. Microsoft wants to make developers happy, so they are showing them the way windows development will work in the future to see how the developers react - what parts they like and don't like. It also means we won't have to wait a year after launch for Longhorn apps to appear.
Yes and NO.... (Score:3, Interesting)
Remember the original version of NT? The version where the GDI was a subsystem onto itself? Back then NT was essentially a micro-kernel approach. However, starting in NT 3.51 the GDI was pulled into the kernel and the result is that NT is less stable in theory.
Of course with enough testing it can be made stable. However, now contrast Windows XP with Windows 2000. Windows 2000 is rock solid, whereas XP can be slow and can be buggy. In essence a step backwards.
Now lets tie this together.
The end result is that applications will not necessarily more stable. In fact instead of being more stable there could be consistent big bad bugs due to bugs in the
Re:Microsoft eliminate blue screen of death... (Score:4, Interesting)
To be fair, when I use WinXP at work, the greatest instability is that whenever Windows Explorer (the file browser) or IE (required for another intranet I need to access) die, they take out the windows desktop and while it sort-of comes back, it'll keep dying if I open another file browser. It lets me save my work, but it doesn't count for long-term stability. But XP usually goes two weeks or so between this.
Re:who cares? (Score:1, Interesting)
Microsoft has not provided adequate tools for updating. This guy does not have a Windows server. So to get his desktops running properly he's going to have to buy Win 2003? Give me a break! This is why people are cheesed off with Windows. Besides, if you knew anything about patches, Automatic Updates, even with SUS, is a bad idea. Updating requires testing, and takes a long time.
If you use the technology fully you have to have Windows servers and buy yet more Microsoft software. That's not acceptable to SMEs, a market Microsoft thinks it's going to get into.
Many people have paid good money to Microsoft for 9x/Me and it's only a few years old. Are you suggesting Microsoft's loyal paying customers should be shot for not upgrading everything every eighteen months? Sounds like a good advert for alternatives to me
Re:Sharing information is a GOOD THING, remember? (Score:2, Interesting)
There are a couple of areas in which they have taken this approach already - the SQLXML extensions to SQLserver 2000 have been upversioned a few times while we wait for Yukon, and various Web Services extensions are also available well in advance of the v2.0 (Whidbey) release of System.Web
Thinking on a bit, if WinFX in Longhorn gives Moft a clean, modular API to replace the spaghetti'd mess that is Win32, it might then become a lot easier to do incremental upgrades to specific areas of the OS functionality.
Re:Wise choice (Score:2, Interesting)
On the one hand, completing Longhorn is a much bigger task, on the other hand a lot of the work's been going on since 2000 at least, and Moft seem to like to dogfood their major releases for a fair while before RTM.
So, we're in late 2003 for the first Technology Preview. From the
I can't imaging that all the slashdotters on earth are collectively as sick to the back teeth of the swiss cheese that is Windows at present as are the Moft people who have to work in the guts of it every day, or try to justify it in the face of entirely reasonable accusations of flakiness, or deal with the support burden, or just try and get their jobs done with the company's products. If the flakiness of Moft products costs many businesses too much, then surely it costs Moft too much in spades.
The very existence of all the Moft internal blogs looks to me like evidence of a much more open and transparent approach and a willingness, a desire even, for as much feedback as possible *before* things are set in stone. Plus it's fun reading that the likes of Don Box and Dare (formerly Carnage4Life of this manor) Obasanjo have been writing their respective bits of Whidbey in Emacs
Bill Gates has a history of 'betting the company'. And as 'Chief Software Architect', this time round the final responsibility for any misdesigns lies clearly and personally with Bill. So I can see how he'd be well in favour of spending *whatever* it costs to finally build a secure, reliable, patchable, maintainable OS on personal as well as commercial grounds.
Re:Not eating their own dogfood? (Score:3, Interesting)
If they're C++ old hands, then they are probably writing it in C++. Yes... C++ is offered as a managed language in .NET. However, I highly doubt that they are writing the core components in managed code.
What they might be doing, though, is re-writing their regular C++ compiler so that it checks for boundry errors and such. They're keeping COM around, since MS's version of .NET needs it. Windows Server 2003 is mostly COM, with .NET sitting on top, which is why it suffers many of the same vulnerabilities that Win2K and WinXP have.
What they will change, however, is the API. They might keep win32 around, deprecating it so that developers won't use it, and just promote pure .NET for development of applications. All those snap-ins, folder explorers, games, notepads, yada, will be written for .NET.
Another thing that bugs me is the insistance that .NET has a VM. It's got an inline compiler that compiles code upon first execution, yes, but subsequent executions are made from the compiled form of the code. Everything would be a bit sluggish at first, but blazing afterwards.
Miguel probably knows more about how it's supposed to work.
Indigo (Score:3, Interesting)
Indigo is really the replacement for COM+, built on top of the web services stack (the WS-* specs). The WS-* specs aim to supplant CORBA as the dominant distributed computing paradigm by enabling any platform to integrate through the various XML protocols. This seems to be the only viable way forward to get true interop between the Windows and ABM (anyone but Microsoft) world.
Some rather interesting things Indigo is trying to do:
- make transactions pervasive in coding, even with volatile objects. Using a "lightweight transaction manager", an in-memory transaction on an ArrayList would take only a microsecond to begin and commit.
- embed the transaction manager in the kernel for durable transactions.
- Provide a set of declarative attributes for setting a service's reliability , transactions, and security, much more flexible and simple than