Slashdot Comment Responses #3 (Your Requests for Open Projects)


One of the questions I asked was “What projects or types of projects should Microsoft look to open up?”

 

I’ll add a HUGE disclaimer here saying that this is not my choice and I have very little say into what we do across the company.  I’m collecting this feedback as a summary for those reading my blog and may have an interest in knowing what people said during this conversation.  It’s interesting information, but don’t expect any of this to suddenly be more open tomorrow.  You should imply no promises or commitments from this post.

 

[Insert any other disclaimer here that might help to ensure that News.Com does not run “Josh Ledgard says ‘Microsoft will Open Up the Longhorn Source” on the front page tomorrow’… doh, they tricked me.]

 

The list is also not a complete one. If you want that you can read through wealth of comments here or on Slashdot. I picked either the most interesting cases or items that had at least 2-3 “votes”. 

 

I’d also like to say that there were several answers to this question that were very well thought out and constructively written from a customer and Microsoft value add perspective. I don’t share all of them here, but I’ve picked a few.  There was a wide range of suggestions for sure, but I think my favorite suggestion is that we provide source code access to Microsoft Bob or other skeleton programs we have lurking around.  I’m now imagining Bob redone in all sorts of Longhorn/Avalon 3d glory.

 

If there is a theme here that I find interesting it is that once you get past the Slashdot discussion of “All of office” the #1 request is for us to open up the code or specifications to things that would make it easier to interoperate in mixed system environments.  This is in addition for us to provide more support for open standards in addition to our own proprietary versions of similar standards.  IMO it wouldn’t seem like too much of a stretch for us to do this with a variety of components.  We offer Windows Services for Unix precisely to make it easier for IT professionals and developers that need to integrate Windows with UNIX and Linux environments.  We’ve classically used this as a business strategy to get into new markets by making the bridge easier for people to cross. Certainly the risk in building the bridge is that you can go the other way as well, but I believe that if you build the bridge then the best platform will win out and I’m very confident that we have the best platform even if you have to pay a bit to play.

 

I’ve heard Steveb explain on several occasions about the additive cost of Windows on a new consumer PC.  Let’s pretend this is $50 dollars.  (Forgive me if these figures aren’t right, but you’ll get the point.) Then assume the average user uses this PC for at least 5 years.  This means that Windows should be providing consumers with at least $10 of value each year over competitors even if you assume the cost of Linux is $0.00 passed to the consumer.

 

I feel pretty good about our ability to provide this added value for consumers, office workers, developers, IT people, etc.  And if our platform is behind for whatever reason… all the more motivation for teams to get a move on and step up the level of competition rather than being complacent.

 

Shell, API’s, or Protocols

There was a call here for not just source code access, but access to more exact specifications or documentation.  This tells me that one of the reasons there are these requests is that we lack a really good documentation or sample story in a lot of cases and people feel the code would serve as the best educational resource that would help them write better applications for our platform.  One interesting suggestion was at least to provide a Shared Source license to this type of code to more ISV partners that feel it would benefit them.  Again, one of the specific requests it would seem is more targeted access to source code through the extension of our shared source initiatives to the people who really need it.

 

Another request here is for Microsoft to be more open with the Explorer.exe code.  Personally I would just like to see us enable people to change their windows themes without having to download a rouge version of a DLL just to change the way windows looks. 

 

The other need for protocol specs is of course to help build better bridges and allow more application interoperability.  Ever try to embed Messenger into an application for example?

 

IE or components of IE

Jonathan: “It’s probably actually costing you money to maintain it (fix security holes etc). Several experts (including the US government CERT I believe) have said to stop using it. Release it under an OSI license and I am sure there are a fair few people out there who are willing to make (and let you guys use/ship) genuine improvements to IE to make it a better browser. Better PNG support. Better W3C HTML standards compliance. Better W3C CSS standards compliance. Better security when it comes to ActiveX and VBScript. And so on.”

 

Jonathan summed up what a lot of people said. I personally feel that this is a backlash against the fact that IE has, until very recently, been something noone sees Microsoft spending any time on.  If we were to continue on the previous trend and not do much to support IE than I might agree with some of the reasons listed above.  As it stands now that there is an IE Team being fairly open with the community I think they should be given a chance to see how much progress they make towards these goals.  Being more open, transparent, trustworthy, and listening to customers does not always have to be about being open source.

 

File formats

·          Any specs to Office file formats presumably to make working between Open Office and MS office easier. One suggestion was at least to support an export to Open Office OASIS standard from MS Office. 

·          Tablet PC Binary Ink format

o         Julian: “Keep the reco algos closed, since I would think that's were most of the IP is... File formats are not the core of intellectual property and they should not be means of protecting a business!”

 

Outdated Applications

·          Travis makes a great plug to get access to 3D Movie Maker.

o         “…I run a site that stores 1344 movies made with this great little program. By letting you use premade actors,props, and backgrounds, they shifted the focus on the voices and story. It's basically a movie prototyping program.
But it still has flaws. It was made right as Direct3D was being released, so it missed the boat. (It's using a software renderer). Some very useful tools were left out (multi-select, texture/model importing, dynamic camera movement) either because they were aiming it at little kids, or because of limitations of the software renderer.
With the source we could make a 3dmm2AVI converter and let others view some of the best 3dmm movies without having to buy/download the program. We've got several C++ programs and a dozen VB coders, we could do so much with this program…

·          Microsoft Bob or other skeletons in our library 😉

·          16-Bit Versions of Windows or windows accessories we no longer ship

 

Developer Platform Tools

I was mostly interested in this category since I work for the Developer Division and thankfully I felt some of the most compelling suggestions fell under this category.  The biggest win I’ve seen for a long time would be the release of the many tools we use internally to support or development.

 

·          Localization Tools: “I suggest Microsoft open sources their localization tools. LocStudio, Helium, you name it. They are all better than the competition's and it would be so great if everybody could use them and not only Microsoft's employees and localization vendors. Also the likelihood of bugs in these tools being fixed would most likely be higher than it is now. ;)”

·          Testing Tools: Mark saidI'm interested on more automated testing tools.” Sure, we’ll be releasing the Team System stuff with Whidbey, but there are still several internal automation platforms, model testing tools, pair wise testing tools, and a ton of smaller, single purpose tools that support software testing.  I don’t have a good reason why these aren’t all already freely available and more open for improvement.  IMO it would be a great PR move for us to release any testing tool that helps us and the industry create more trustworthy software.  Sure, one might think we would lose some of our “best practices” to competitors, but I want to be really serious about improving the quality of software across the industry and releasing these tools would be a good step in the right direction.

·          Breaking the VSS Tie: This was a strange one. It seemed a lot of people didn’t know that we already have a way to add any source control mechanism into Visual Studio.  Anyone could build a source control integration package that makes VS work with CVS, Sourceforge, Gotdotnet, etc and many of them exist today.

·          Compact Framework: Chuck “This is a longshot, but the Compact Framework. The CF is woefully inadequate and doesn't play as well as everyone would like with the desktop version of the framework. You cannot reasonably develop a .NET CF application without using a plethora of other 3rd party libraries. 3rd party libraries are great, they help move the development market - except all the 3rd party libraries I use are open source already. The OpenNETCF project isn't moving as fast as it could; an open source CF sponsored by MS would be a godsend - of course it would have to be on SourceForge and not GDN.”

o         Good idea, we should probably do more to get behind the opennetcf efforts and release more components through this model.

·          “Wiki-Ize MSDN”: A couple of people mentioned this concept. There are also people internally that would like to see us move closer to this model. Today you can see community annotations to the Longhorn SDK documentation, but we should probably go further here.   I can’t help but look at sites like www.pinvoke.net and wonder why we don’t have that type of mechanism in place for every framework class. We have a hard time documenting and providing samples for everything, so why not let everyone contribute to shared knowledge. 

·          Web Service Utilities: “Some of the web-service related utilities. For example, xsd.exe, wsdl.exe (& wsewsdl2.exe) and WebServiceStudio. This would allow developers to modify them to meet their specific needs and would also allow the tools to become more robust.”

·          All or Pieces of Visual Studio: Jonathan I have wondered before if Microsoft could make some of the non-core parts of Visual Studio open source. For example, all the functionality of the Solution Explorer window could be implemented through the publicly documented automation interface so perhaps it would be possible to release it as a mini open source project. If there were mini-projects for things like the Solution Explorer and the Output Window and perhaps even the text editor users could fix parts of the program that annoys them and help make Visual Studio better.”

o         Another good idea, but no promises.

·          Developer Other

o         Build Tools

o         C/C++ Runtime

o         .Net Framework and Microsoft Business Framework

o         VS.Net Command Line and External Tools

 

Other

·          NTFS

·          Microsoft ACPI code

·          Com/Dcom/ActiveX

·          IIS 5.0 and 6.0 (older versions especially to assist with security support)

·          Window CE 5.0

·          Activesynch: Peter “Again, it's not core, it's got some real simple problems that would really benefit from OS work (like, how about using link-local addresses instead of 192.168.55 which happened to conflict with one of our internal networks and broke Activesync until we renumbered), and it would improve market share for the Pocket PC if it could be ported to other operating systems.”

·          DirectX/D3D: “The way I see it, MS would still have their proprietary XNA technology. Therefore there's no harm in opening up D3D, which would really benefit and bring relief to many 3D software developers and game programmers.
While you're doing that it might be handy to open up DirectInput along side this. Almost everyone uses this as the defacto method of grabbing high resolution input and handling many devices.”

·          Unsupported Device Firmware

 

I’ll repeat my disclaimer here…

I’ll add a HUGE disclaimer here saying that this is not my choice and I have very little say into what we do across the company.  I’m collecting this feedback as a summary for those reading my blog and may have an interest in knowing what people said during this conversation.  It’s interesting information, but don’t expect any of this to suddenly be more open tomorrow.  You should imply no promises or commitments from this post.

Comments (30)
  1. David Cumps says:

    "Personally I would just like to see us enable people to change their windows themes without having to download a rouge version of a DLL just to change the way windows looks." Amen! I don’t like the fact that I have to patch uxtheme.dll to get my favorite Watercolor theme to work on XP and 2003. What’s the reason behind restricting themes anyway?

    About the IE team being open, I believe they will get the deserved respect when people are noticing changes in the support area (png, css, etc), the biggest comment I read about the IE team being more open is "who cares if they talk with us, it’s actions that count". And while I don’t really die from the standards of IE, I tend to agree with their statement, blogging and telling about how much work you do is one thing, releasing things would quickly change the attitude towards the IE team in a possitive manner (imagine what a community feedback you’d get if suddenly with a patch IE has all the PNG support everyobdy asks for, and CSS support, IE would be praised) (and I’m sure there will be a way for admins to set IE to the ‘old-style’ mode for their applications who were designed with the weird CSS handling.)

  2. josh ledgard says:

    David: I don’t know the reason other than potentially trying to ensure that Out of Box windows user expeirinces are all the same and there isn’t a different between going from a Gateway PC to an HP PC. I do see merrit to that, but on the other hand I’m someone who gets bored with the UI every few months.

    Regarding the openness. It takes time. We started talking to people about VS seriously 2-3 years ago, but most of the "actions" won’t be seen until we release whidbey.

  3. ShadowChaser says:

    I just wanted to comment on this item:

    "Breaking the VSS Tie: This was a strange one. It seemed a lot of people didn’t know that we already have a way to add any source control mechanism into Visual Studio. Anyone could build a source control integration package that makes VS work with CVS, Sourceforge, Gotdotnet, etc and many of them exist today. "

    You’re right – however, Microsoft FORCES developers to sign agreements when licensing/using the SCC interfaces! It’s literally impossible for open source applications such as CVS to hook into Visual Studio – the Microsoft non disclosure license is incompatible with open source.

    Why is there a non disclosure agreement on SCC? 🙁 🙁 🙁

  4. Mike Dimmick says:

    Specifying NTFS fully would assist in allowing other OSes to read and write to volumes originally written by NT/2000/XP, but there might be some mileage in adding ext2 support to Services for Unix – if there’s a way to do this without getting GPL-infected.

    Compact Framework’s major problem in version 1.0 was that the Program Managers responsible ripped out almost every API with any power. I understand the reasoning for reducing the overload set – more overloads = more code + more metadata, increasing the size of the binaries – but they removed the overloads with many arguments rather than those with few. This leaves NETCF 1.0 neutered. Example: the only version of WaitOne remaining is the version which blocks forever – the overload taking a timeout value has been removed.

    To make matters worse, COM interop was removed, as was much of the ‘flat’ API marshalling power. This makes interoperating with existing code of more than trivial complexity a total nightmare.

    Recently we’ve been dealing with window ownership problems. In NETCF’s Windows Forms, it’s not possible to create a new form owned by an existing one – all forms are created as top-level windows. If the user uses the Memory control panel’s Running Programs tab (on Pocket PC) they can switch to a form that’s not at the front – if each form was owned by the previous form in the stack this would not happen, as Windows brings owned forms in front of their owner.

    Apparently CF 2.0 has been rethought and many of the features removed from 1.0 have been reintroduced – including COM interop and marshalling.

  5. Rob says:

    Josh, awesome, awesome work! Keep it up!

    rob

  6. Mike Dimmick wrote:

    "Specifying NTFS fully would assist in allowing other OSes to read and write to volumes originally written by NT/2000/XP, but there might be some mileage in adding ext2 support to Services for Unix – if there’s a way to do this without getting GPL-infected."

    No changes in SFU should be necessary to support ext2. What’s needed is to get the IFS kit (see http://www.microsoft.com/whdc/DevTools/IFSKit/default.mspx). But it sucks bigtime that it costs 899$ to get it. WHY?!? Would be a lot better if this was just as free to get as the Platform SDK and all the other SDK’s Microsoft is providing. All(?) would benefit with broader support for filesystems in Windows.

    Related to this is that the Windows DDK isn’t available for download anymore. You can order a CD for s/h costs, but why can’t we just download an ISO? Since I don’t really need the DDK I haven’t ordered it (yet), but occasionally it would be nice to have the header-files there handy to check out some definitions etc.

    To check out the sources of Microsoft Bob would of course be nice, but I’d be much happier if I can download any of these kits/DDKs for free. And opening the source for some product will involve some management, but to make the items mentioned above available for download would only need you to create the ISOs and publish them on http://www.microsoft.com/downloads 🙂

  7. AT says:

    Andreas Häber: Do you know a lot of people who will wrote own filesystems for Windows ? How much time do you expected they will spend on writing new FS ? How much this time will cost them ? Now compare this cost with 899USD per developer.

    DDK is not longer distributed at S/H costs. It cost 199USD (but it can be downloaded for "free" if you have MSDN subscription). Same questions again. Are you willing hobbist developers wrote buggy kernel mode drivers for you ?

  8. Stephane Rodriguez says:

    jl wrote : "Any specs to Office file formats presumably to make working between Open Office and MS office easier"

    I don’t get it. Every single byte of the Office file formats was documented in MSDN, under the Office 95 and Office 97 topics. And they got removed from MSDN for the same reason that a bunch of C++ APIs are being removed with every MSDN release. Last MSDN with such topics was MSDN Oct 1998 I believe. Want to make us happy?

    – republish all MSJ archives : today a bunch of them either are not accessible anymore, or are 404 pointers. Worth every single penny.

    – identify and republish a MSDN classic DVD set. Worth every single penny.

  9. What about the custom components in Office and Visual Studio (e.g. the menus, dockable windows, smart tags, completion lists like those used in Intellisense)? In this case, we wouldn’t even need the source, just access to the components. Just sharing them with the rest of the Microsoft teams would be nice so that we could get a consistent experience across applications.

  10. Jonathan Wilson says:

    the thing to remember is that Microsoft is only going to be more open if being more open has no negatives for them.

    For example, open sourcing the C++ compiler would be a negative for them since then everybody learns the secrets of what makes their compiler good (or not good)

    Giving out the specs for NTFS (although from what I have seen, just having some kind of spec document wont necessarily make NTFS on linux magically start working better) would also be a negative.

    On the other hand, opening the APIs used to integrate source-code-control-systems with Visual Studio is a positive.

    Microsoft makes no money whatsoever off SCC systems (it only makes SourceSafe and it gives that free with Visual Studio anyway).

    Releasing the source code to IIS would also probobly be a positive. As far as I know, Microsoft makes no money from IIS (I dont think they even do much development on it) and with most of the web (small sites in particular) switching to apache because of the security issues with IIS, open-sourcing it would probobly be a good move.

    Documenting the stuff in dlls like shell32.dll would probobly be a negative. It would enable shell extentions (and other programs that want to talk to the shell) to use those functions. But it would force microsoft to be backwards compatible with 3rd party apps that use these newly documented shell interfaces.

    That said, microsoft has documented more functions in shell32.dll recently (a fair few were documented in response to the USDOJ lawsuit).

    Oh and something related to this whole "open" thing is this comment from a MS employee on microsoft.public.vstudio.development:

    http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&selm=oMoJFXNlDHA.2624%40cpmsftngxa06.phx.gbl

    indicating that they were planning to release more of the C Runtime Source at some point in the future. Would be good to see this actually happen… 🙂

  11. Jim says:

    > I personally feel that this is a backlash against the fact that IE has, until very recently, been something noone sees Microsoft spending any time on. If we were to continue on the previous trend and not do much to support IE than I might agree with some of the reasons listed above. As it stands now that there is an IE Team being fairly open with the community I think they should be given a chance to see how much progress they make towards these goals.

    The thing is, the official story from Microsoft, that the IE team has yet to contradict, is that there won’t be any new versions of Internet Explorer. Ever. The next web browsing-type application is going to be the Longhorn shell, and tough luck to everybody not using Longhorn, tough luck to the web developers that have to support IE 6 for another six or seven years, and tough luck to the businesses that have to pay for the extra time needed for IE workarounds.

    The IE team can blog all they want, but until some progress is made with support for HTML, CSS, PNG, etc, it will be utterly meaningless, and with every passing day, the backlash just gets worse. I was personally very hopeful when the IE team started blogging – but after reading all about such gems as "how to launch favorites", and reading nothing about the next version, I’ve become even more disillusioned and bitter towards IE. The blog has backfired by getting people’s hopes up and then dashing them by ignoring everybody clamouring for proper CSS support. All anybody hears is "we can’t promise, we can’t promise". You promised PNG support way back in IE 4.0… the time for promises is long past – now’s the time for action.

  12. Digging .NET says:

    Microsoft and open source

  13. AT: I don’t know about people who wants to write FS drivers for Windows, but I do know that a lot of people who dual boot Linux and Windows would like to be able to access ext2 and other filesystems from Windows, instead of relying on FAT for interop between those OSs. See [1] and [2] for some examples.

    Also I believe there would be some academic interest in creating new filesystems for Windows.

    And yeah, there are already sooo many incredible buggy drivers around for Windows from hw vendors etc already, so what does it matter by making the DDK downloadable?

    Like I wrote earlier there are some definitions in the header files there which are interesting for developers in userland too.

    If someone writes a buggy driver then I don’t have to install it 😉 but why keep it as this big secret thing, forcing people to use reverse engineering techniques instead (see [2]) which WILL create more buggy drivers.

    [1] http://uranus.it.swin.edu.au/~jn/linux/ext2ifs.htm

    [2] http://www.acc.umu.se/~bosse/

  14. AT says:

    Andreas:

    You see problem.

    I see solution.

    Dual boot is the past.

    Use http://www.vmware.com/, http://plex86.sourceforge.net/, http://www.microsoft.com/windowsserversystem/virtualserver/ or http://bochs.sourceforge.net/ (if you are using supercomputer 😉

    Configure minimal *nux with Samba, FTP or NFS (with Windows SFU). It will boot in less then 5-10Mb of memory inside Virtual PC and you will be able to access all your files in read/write mode !!!

    This way you will not damage your native file systems.

    As well you will have all available file systems supported – not only EXT2 or EXT3, but ReiserFS, XFS, JFS!!

    As for performance – I think that decrease will not be that critical for your usage. Sure, aggressive windows file cache will not be used and network access will have some additional overhead – but I think that you are not running production server on dual-boot or recompiling linux kernel sources using Visual Studio from EXT2 file system 😉

    P.S> As you already pointed to me – there is ntifs.h file with a lot of function created _without_ assistance from Microsoft. So? What the problem? Or you are waiting for samples and guide from Microsoft created "How to wrote own file system"? This guide already exists – but it cost a few! Microsoft not using "Paypal Donate" link – they use link to http://shop.microsoft.com web-site for this.

    As for FileSystems in research – I believe that Microsoft already provide complete Windows source for university researchers ( http://research.microsoft.com/collaboration/university/source.aspx ). As well you are doing real for-profit research – 800 USD is very little cost. Latest notebook from HP/Dell costs twice more!

    I think that it’s even possible to receive a Microsoft grant for Windows research!

    I still do not any real benefits from making DDK and IFS "free".

    I believe that Microsoft doing everything correctly here.

  15. nickos says:

    "Another request here is for Microsoft to be more open with the Explorer.exe code. Personally I would just like to see us enable people to change their windows themes without having to download a rouge version of a DLL just to change the way windows looks."

    I saw some requests on Slashdot to go further than opening up explorer.exe and open up the window manager part of the GUI. There’s already many different explorer.exe alternatives out there – I want to be able to change the fundamental behaviour of windows (raise-on-focus for example). This would also allow users to have more control over the look of their system than any theming engine could allow.

  16. josh ledgard says:

    Nickos: But really the user request is all about enabling easy/deep customizations. There would be other ways to enable this and make it easier for peopel who don’t want to deal with source code even.

  17. Norman Diamond says:

    8/31/2004 5:36 AM Andreas Häber

    > If someone writes a buggy driver then I

    > don’t have to install it 😉

    If it came from Microsoft and was built into the Windows CD then you DO have to install it. Opening up the driver source would yield a chance of getting these drivers fixed.

  18. nickos says:

    Thanks for replying Josh,

    With respect, I disagree. The FVWM window manager for X can be customised in thousands of ways, and yet there are still hundreds of other window managers out there. In other words, a more customisable "window manager" in Windows will not satisfy everyone. At least in X land the window manager is a seperate process that can be swapped out with an alternative, open source or not.

    There will always be a features that the developers of a closed source application never think of. For example, where I work we have to use a certain (closed source) text editor. It’s pretty good but lacks (a sorely missed) feature that would probably take me an hour or two to implement if I had access to the code. Open source allows coders like you, me and most of the readers of your blog to customise their tools in ways the original developers never thought of.

    By all means provide a deeply customisable GUI, but let others replace it easily (and give them the info so that they can write replacements).

  19. tuco says:

    >"…I’m very confident that we have the best >platform even if you have to pay a bit to >play…."

    The EULA is part of that "platform". And, sorry, the MS EULA is so bad it brings the OS down with it. Read the MS EULA please. And read the Home Recording Act and the Fair Use Act.

    I think it is easy to forget when you work at MS and can get ALL their software for a few buck (if not free) that the rest of us can’t. For example, you want to put an OS on your desktop and laptop and have to buy two licenses! This is a debatable issue with the Fair Use Act.

    And granting Microsoft the right, by accepting their EULA, to enter and make changes to your computer without asking you! C’mon, you don’t let Sears enter your home at anytime because you bought a range from them.

    Plus excluding the user to rights of "quiet possession". Imagine if your landlord had that right. You’d be alarmed and have him/her in court immediately.

    The list goes on and I realize you can’t do anything about that. But use your words of "the best" carefully, please. The "best" in software is not always measured by how functional it is but what it will let you do. And MS has to have the most restricted EULA ever produced.

    I think the MS law department should be reading your blog.

    My two-cents….

  20. Josh Ledgard says:

    Nickos: I guess I was focussed more on what I would consider the 80% case. People like me who simply want to change the looks more often rather than people who want to drasticly change windowing behavior and would go so far as to dig into the code to do it. Would I like the both worlds… probably… would the magority of windows users want to be confused by the different schemes that would emerge??? I don’t know. Would more windows users and OEMs like to customize the way things look in a more meaningfull way… probably. It’s always going to be a time constraint. I’m willing to be that customizibility is the first thing cut when it comes down to scaling back a deliverable to reduce the testing matrix. It’s no easy task to write software used by such a wide distribution of machine configurations. Even with VS, there have been instances where our UI suffered bugs from display drivers written incorrectly. (We are working to make sure we look good under non-standard windows themes though. 🙂 )

  21. josh ledgard says:

    tuco: I’m not a lawyer, nor do i really want to get into the business of playing one in this space. I have, however, since read the eula. I can say that I see the reason why such wide terms end up making there way into the documents. Specifically the example you mention. I believe that the Eula has to cover any scenarios delivered by the product rather than on a case by case bases. So, in this case, the Eula is probably written that way to cover the automatic windows update scenarios. Sure, you could turn it off and only do updates when you have allowed them, but the functionality is there for automatic updates so it is covered in the eula rather than having people agree to new terms every time they turn on another feature. Again, I don’t know the full reasons, I’m not a laywer, and am simply guessing here.

    Regarding pricing. Honestly I’ll concede that I have been caught off guard several times in the last year by the prices of things like Visual Studio in the wild and I do get caught up in the trap of "hey, I can just download it or pay for deep discounts at the company store." So yes, I’ll wager to say my view is probably colored. I’m excited, however, about some of the changes being made to the VS Skus… namely the express versions. Sure, you won’t get everything, but for a hobbyist getting started they are pretty darn powerful IDE‘s.

  22. AT: Sorry, but I think you misunderstand what I’m saying. The problem here is that by making the DDK etc. available for free download will make it easier for people to create support for various stuff (such as fs-drivers but not limited to that) and also to get a better understanding of Windows internals when writing user-level software (two reasons of the top of my head: 1) check out the header files which contains valuable information. 2) get better information for creating INF-based setups the DDK provides more information then the SDK does. Happily the DDK docs is available in MSDN Library now. And yeah, I *KNOW* there are other ways to write setup programs, but that discussion is way out of context here…).

    I am not convinced by your arguments that it is better to have us pay lot’s of money to get access to the DDK. It will be sad if the Platform SDK and the .NET SDK also starts to cost money :/

    Norman Diamond:

    That was way out of context.

    1) I was talking about drivers written by 3rd parties, especially "hobby"-programmers creating filesystem drivers etc. I doubt that Microsoft will ship Windows with for example ext3 drivers written by some "hobby"-programmer…

    2) To get a driver on the Windows CD it has to pass a certain level of quality.

    3) I agree that it would be kind of them to provide the source code of their drivers to use as examples etc. But giving away such inside information will also make it a lot easier for developers to write software which isn’t backwards compatible (read: relying on something internal in a driver, which then breaks in the next servicepack because the internals was changed…)

  23. AT says:

    Andreas: You will never imagine how many hard-disks data I’ve to restore because of NT4 FAT32 from Mark Russinovich. And I’m think Mark is not "hobbyst"-programmer.

    People tend to install any software they find in Internet. User-mode software does not cause real problems if crashed for NT-family, real problems are with Win9x family only.

    But kernel-mode software in NT is software with high cost for any bug.

    BSOD, data corruption, security compromise, etc … Kernel mode software effect entire system.

    So I preffer only people who can spend 200 USD for 18 CD kit with bunch of driver-development information ( http://www.microsoft.com/whdc/devtools/ddk/suite/default.mspx ) wrote drivers for users.

    Even more – if you need simply a documentation – it’s accessible for free on http://msdn.microsoft.com/library/en-us/gstart/hh/gstart/Z_gstart_HDR_8283b593-7838-4f10-9c4c-300807bae95f.xml.asp URL 😉

    P.S> As for certification to include your drivers on Windows CDs – this is easy and cheap. ( http://www.microsoft.com/whdc/whql/policies/testing.mspx )

  24. tuco says:

    “…I’m excited, however, about some of the changes being made to the VS Skus… namely the express versions. Sure, you won’t get everything, but for a hobbyist getting started they are pretty darn powerful IDE’s …”

    Yes, Josh, Microsoft’s IDE’s are powerful and well integrated indeed. And it is cited often for reasons why MS is a good development platform. No arguments there.

    But there are times when something is good enough even though it is not the best. I think of a car, for example. How many of us drive the very best car around? Some of us have to manually pull a lever to adjust the seat and some can push a button.

    For me, it is the same with software. I don’t always need the best. If I can get a bunch of Open Source development tools that may be less than integrated but get the job done without a ton of restrictions, I feel that is way to go. Especially since the price is right and diffidently good for the hobbyist.

    For instance, I wanted to learn more about setting up a web server, a proxy server, a SQL database and some other things (just to understand it more and testing). With my spare old hardware lying around, I set up a network and did it with all Open Source stuff for cheap. I mean really cheap. I estimated (not counting “lite” versions of MS stuff) it would have cost approximately $15,000 to get MS equivalent set up ( depends, of course, how you estimate the number of users but the OSS had no user restrictions). So with MS you have to pay more than a bit to play and learn.

    Another example is at work I’m issued a Windows box. At home, I have Linux, BSD, and a Mac. Now I was writing a small utility at work and I used Python. I worked on it at home on my Linux box as well as at work on my Windows box and on the road with my PowerBook. The multiplatform nature was nice. I don’t even know what it would take to do multiplatform work with a Microsoft product. I imagine it can be done (I’m not familiar with all of MS tools). Maybe this is the direction MS can focus on: multiplatform without trying to lock the user into their products and fewer strings attached.

    Thanx for the reply.

  25. Travis Wells says:

    Wow.

    I realize that it’s not up to you, but your mention of 3D Movie Maker just made a lot of 3DMMers happy!

Comments are closed.

Skip to main content