In a move that shows just how far Microsoft has come, and how pervasive open-source software is in certain areas, the software powerhouse is, for the first time, including open-source technology in one of its shipping products. Microsoft plans to include the Message Passing Interface in its Windows Server 2003 Compute Cluster Edition, which went to public beta this week at the Microsoft Developers Conference here and is on track to ship in the first half of next year.
This is not the first time Microsoft have shipped Open Source technology. They shipped the GNU compiler suite in Interix, which became their Services for Unix. They carefully followed the GPL in doing so.
IMHO, the open source community seems to do a lot better with codecs, APIs, etc. thatn it does on UI’s and such. I wouldn’t mind at all having all open documents and APIs with proprietary ‘shell’ apps wrapped around them.
… like OSX? 😉
The problem is that a lot of open source developers, usually on the smaller projects, don’t understand the concept of seperating the working code from the interface code. The working codes gets tied in with the interface code thereby making it very difficult for designer types to fix the interface.
but, Firefox does a really good at the seperation.
Also, I think the IP stack for NT was taken from a BSD-liscensed project. I’m sure they’ve used other OSS as well.
Another example (please correct me if I’m wrong) is the TCP/IP stack in Microsoft’s Windows products being BSD licensed code – I’m guessing this counts as open source technology in a shipped product?
BSD TCP-Stack in windows, anyone?
4.2BSD IP stack, zlib, kerberos, openSSL, etc, etc. This is a completely useless article.
http://lists.matrixlist.com/pipermail/pc_support/2005-September/000…
Damien
re not the first time, I have to agree that using the BSD TCP/IP stack was probably a more significant move…
… of news.
The article was interessting though, in regard to the implementation. It sure did sound like MS is going to enhance and embrace this one, too.
However, there’s been a certain tendency in recent time for MS to actually participate in opensource projects. I’ll leave to you, what to make of that – but it is interesting.
The huge succes for F/LOSS in Europe will probably ‘force’ MS to support open standards 100% correctly in due time. Let’s hope that, because it means we can all share information without having to use the same platform. And that means freedom for us and that’s the idea
dylansmrjones
kristian AT herkild DOT dk
You would expect better from them. EVERYBODY knows that Microsoft had a BSD TCP/IP in windows at one time.
even the ftp client in Windows 95 onwards is from BSD; and nearly all of SFU has seen OpenBSD code at some time
This is so far from the first product they’ve shipped with OSS code its not even an excusable mistake to make. As someone said above, they’ve even used GPL code (legally)
wasn’t internet explorer derived from open source software? from mosaic right?
wasn’t internet explorer derived from open source software? from mosaic right?
Sure it wasn’t based on spyglass?
Not sure if your kidding about your statement:
“Sure it wasn’t based on spyglass?”
IE is or was Spyglass
From the IE about box:
Based on NCSA Mosaic. NCSA Mosaic(TM); was developed at the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign.
Distributed under a licensing agreement with Spyglass, Inc.
NCSA Mosiac is sort of open source:
The University of Illinois at Urbana-Champaign has made the source code for the Macintosh and Microsoft Windows versions of NCSA Mosaic available free of charge ONLY TO ACADEMIC INSTITUTIONS (UNIVERSITY LEVEL) AND UNITED STATES GOVERNMENT AGENCIES. Applications from academic institutions must be signed by a DEPARTMENT HEAD OR HIGHER and applications by government agencies must be signed by someone at the CONTRACTING OFFICER LEVEL OR HIGHER.
From: http://archive.ncsa.uiuc.edu/SDG/Software/Mosaic/License/LicenseApp…
… the first virus to find its way into the Windows operating system.
If they didn’t include MPI for clustering software they’d be laughed at by literally everyone. They could try to push their own stuff, but in the end: Simulation code doesn’t want to be rewritten. Scientific code has a tendency to be *very* badly maintained. And the people who write it tend to not want to learn new technologies often…
They’d be wasting their time without MPI.
well, noone is laughing at their “me too” versions of pdf or flash, and those are widely supported standards.
honestly though, i wish they would do more of this instead of their normal subversion strategy.
Oh the irony. I guess let’s put this on the list next to using NetBSD for their DNS servers.
Asked by eWEEK what Microsoft will give back to the open-source community for the MPI component, which is licensed under the BSD and not the GNU General Public License (GPL), Faenov said all fixes will be given back, while “we’ll probably give the changes back as well.”
That was the shocking part.
No. The are releasing this as a way to:
1) Improve their code.
2) Spread business via:
a) OSS
b) Get into the entrenced world of clusters. Clustering software has strong roots in *nix. MS is the new kid on the block when it comes to clustering.
That’s really the reason I don’t care for Microsoft. I believe that their software is improving by leaps and bounds, (hopefully I can tone down a lot of the “interface innovation” in Vista.) but their hostility toward open standards has always bothered me. If they continue to make progress in the interoperability area, they may one day find me as one of their customers (not that I’ll give up my ibook.) Open source software has its advantages, and has made a very positive contribution to software as we know it today, but it is not the entire universe. Right tool for the right job, I say, and if Microsoft products start to work well outside of a non-Microsoft universe, then they will get my money. Apple, on the other hand, continues to be a pain in the ass no matter how much I like their software. Examples include: DRM on AAC, locking out of 3rd party DVD burners on older versions of iDVD, etc. However, it seems Jobs is becoming more lax in his control freak issues over MacOS X now that he has his new pet, the iPod.
I guess what I’m trying to say is that I like companies that support standards, and if they have to change them to move forward, they submit their changes to a standards body so everyone can benefit. (Also, if everyone follows standards, it’s less of a hassle for the end user.)
can you imagine the RIAA going for itms without drm? its like the HD-DVD bs on vista. your going to need a monitor that doesnt exist yet, and a vid card with an internal bus capable of 3gb/s transfer rate in both directions because frames may go back and forth up to fifteen times to be totally decrypted.
honestly, i think its up to the market to just say no to drm.
It’s just that the open infrastructure is a requirement for competition and competition isn’t that good for most companies. So if you have the power to kill the open infrastructure you can make a living selling a closed one.
“Portions of this product are based in part on the work of the Regents of the University of California, Berkeley and its contributors. Because Microsoft has included the Regents of the University of California, Berkeley, software in this product, Microsoft is required to include the following text that accompanied such software:”
Just thoght i would put in a quick quote and link in regards to BSD:
http://www.microsoft.com/technet/prodtechnol/windowsserver2003/libr…
“Oh the irony. I guess let’s put this on the list next to using NetBSD for their DNS servers.”
They did not use BSD, they’re hosting company did (does).
LIES!!!111 >_<
roflmaomg
As ma_d mentioned, MPI is ubiquitous in high-performance computing, so it is a must for any serious contender in the clustering space. MPICH has long been the reference implementation. It (and it’s reimplementation MPICH2) have been BSD licensed from the initial release. For this reason, they have been the basis of many, if not most, vendor releases of MPI. It is certainly the basis for the MPI the IBM’s Blue Gene system, Cray’s Red Storm/Strider and X1 systems, and many earlier systems. Many network vendor MPI implementations (like Myrinet and Infiniband) also use MPICH as the basis.
No doubt the license is one (among many other) reasons for this; there is literally no downside for vendors looking for an MPI implementation that takes them 90% of the way.
…and they are shipping it only 5 years after I used and abused it in undergrad parallel processing courses.
I suppose when you are late to the table, you’ll take anything you can get and claim it’s the best thing since sliced bread.
nt4 resource kit had perl, vi, and a few other open source products.
Microsoft loves opensource, they just don’t like the GNU GPL.
– Jesse McNelis
This is Microsoft’s tactic in every field that they enter. They’re the newbies on the block in cluster computing. What they do is go in, interoperate with any standards that exist. As soon as they garner a large share of that market, they start adding proprietary features that make Windows a dead-end for projects running on their platform.
Look at OpenGL. They played nice for awhile, and then made DirectX. Office file-format compatibility back in the mid 90s. TCP/IPX and basic networking – then came their file sharing, Active Directory, etc. Early mail support was standards based (IMAP/Pop), now we have Exchange and whatever protocol it uses. Internet Explorer started out playing nice, then added ActiveX and all kinds of CSS compliance issues. Netscape died, so did IE development. Firefox causes a stir, IE development starts again.
This shouldn’t surprise anybody. (Also read everybody else’s comments about other Open Source components they’ve used).
Indeed, IE has been derived from Spyglass. An interesting detail is that by bundling IE with windows, Microsoft saved a lot of money:
http://en.wikipedia.org/wiki/Spyglass
“The deal stipulated that Spyglass would receive a base quarterly fee for the Mosiac license plus a royalty from Microsoft’s Internet Explorer revenue.
Microsoft subsequently bundled Internet Explorer with Windows, and thus (making no direct revenues on IE) paid only the minimum quarterly fee. In 1997, Spyglass threatened Microsoft with a contractual audit, in response to which Microsoft settled for US$8 million.”
lol..m$ and OSS are two immiscible liquids.. :->