The Linux source might be GPL but at the end of the day if Linus or his advisors don't like the changes you suggest it won't get into the official kernel.
True, but many Linux vendors don't ship an 'official' kernel and instead fork it with add-ons that Linus hasn't gotten around to including yet. (e.g. ISDN and NFS fixes)
One thing I have noticed reading linux-kernel is that the development process is only 'open' to those coming with working code. That is, bug reports and feature requests are usually ignored unless there is a patch attached to the message. (Fine, because linux-kernel is about engineering, not marketing.) Commercial vendors, on the other hand, have mechinisms in place for bug reports and 'enhancement requests'. However, such reports typically disappear into a black hole from a user perspective - there is very little feedback until a fixlist emerges for the new version.
Sun is embarking on an opportunity take the advantages of both open source and commercial development practices. However, if Solaris patches disappear into the same black hole as enhancement requests, it's not going to work.
I would hope that (1) Sun starts an open list, similar to linux-kernel, to discuss Solaris development issues and review patches, and (2) there is an open central repository where users can download submitted patches. (Similar to the way Linux users get access to 'unapproved' stuff like ISDN or NFS.) Even if someone can't redistribute "MacMillian Solaris", if users have to wait for Sun's blessing to get access to patches, it's not going to work.
I have my doubts that Winblows95/98 is an example of good software engineering, its more like a HACK job, or would that be better said as ONE_BIG_KLUDGE?;-)
Don't forget that the #1 design goal of Windows 95 was to be backwards compatible with 10 years of hardware drivers and DOS/Win16 programs. They got that part right.
The #2 goal was to provide a similar API to Windows NT so that at some point in the future they could obsolete DOS/Windows. They got that mostly right, although 9x makes so much money that DOSWin isn't extinct yet.
Later on, with DirectX, they wanted to provide a better game environment than DOS could provide. I haven't seen too many DOS games lately, so I assume that's going well.
Note that security or reliablity were never high up on the design goal list for Win 9x. They already were selling Windows NT, you know, but even to this day people choose backwards compatiblity and games over a superior OS.
If this is the case then sun should stop deceiving people into thinking that this is open source. Whenever Sun talks to the press...
You are going to have to provide a reference to prove that - I couldn't find one.
Sun's standard PR bit on SCSL says:
The CSL model is the result of Sun's continued commitment to the open development of key technologies, and is based on Sun's community source principles, which include immediate open access, increased innovation, faster commercialization, and access for students.
ZDNet (referenced by/. a few days ago) makes the comparision with Linux, but they're ZDNet, not Sun. They quote someone from Sun:
At the end of the day, the issue is how you allow innovation, but also have a reasonable process by which the community sticks together on the core as it evolves," said Anil Gadre, general manager of Sun's Solaris division.
Neither statement looks like an appeal for the support from Open Source (TM) advocates or developers. AFAIK, they didn't enlist Eric Raymond or make any appeals to Linux developers.
And, they will probably get some free labor, although it will be from commercial developers and hardware manufacturers, not the OSS crowd. Which makes sense, because by-in-large Sun doesn't make any money from the OSS crowd, so why should they care?
While Sun is allowed to derive benefit from the community, the community is placed in a state of perpetual legal risk
Did it occur to you that perhaps the "community" referred to in the Community Source Licence is not the same community that comprises the Linux/GNU user base? Just because they mention "community" and "source", there is no reason to make the egotistical leap that they are talking about the "Open Source Community".
There is a very large, and very valid community that can benefit from Sun's SCSL. Namely people who depend on Sun products and don't give a shit about redistribution or forking or winning the hearts and minds of Linux users. If that's not you, why should you care? Source for Solaris can mean bugfixes today and not when Sun gets around to it, allows more code optimization, makes it easier to write device drivers, and so on.
(Note - Same argument goes for those who flame Apple's source licence.)
Furthermore, you appear to be demanding that Sun should start giving away commercial software simply because it would be useful to you and others. Why pick on Sun's SCSL chart module? Why not demand the dozens of other commercial closed source chart programs (like MS Excel)? Are you trying to flame Sun until they GPL all of their software just because the SCSL makes them seem amenable?
You might be right. There are already NT4 "Post-SP6" hotfixes out, and SP6 hasn't even been released yet.
Re:auto-attenuating Harmon Kardon subwoofer
on
New iMac Rolled Out
·
· Score: 1
Isn't this argument over software subwoofer detection rather lame? After all, this problem was solved in a "plug and play" manner in hardware 40 years ago. (Or was someone too cheap to solder in a $.10headphone jack?)
It's kind of interesting that none of the commercial vendors have ponied up the money for a TPC benchmark on Linux. I would think that it's a given that Oracle/Linux beats Oracle/NT and probably MSSQL/NT as well....
Perhaps RedHat (et al) are too busy chasing the "workgroup server" and small business markets to bother with trying to sell database servers.
I recall 1994, when a NT administrator could demand 25-50% more money than a NetWare administrator. As NT's usage spread, and Microsoft was successful in promoting it's education programs, the price of NT administrators dropped through the floor.
UNIX (and AS/400, VMS, etc) admins are expensive because getting admin access to a Unix system is difficult. so fewer people learn. A vicious catch-22.
Linux, on the other hand doesn't have the access problem. I'm sure that in a couple years there wil be enough Linux admins that labor costs will be comparable to other PC NOSes.
2) Remove all private interfaces between the operating system development group and all other software development groups at Microsoft
You are hitting at the heart of the integrated product strategy at Microsoft. Consider the introduction of OLE2 in 1993. Microsoft released the operating system feature (OLE) as part of MS Office, and as a tool in VisualBasic. Users instantly had an example of the functionality and could go out and implement it themselves.
Compare this to Apple OpenDoc. Apple released the API, but refused to 'integrate' it in ClarisWorks or the MacOS Finder. Result was that nobody really understood the possibilities and very few applications ever bothered with the API.
If a brick wall is instituted between the MS OS group and the Apps group, the result is that most new OS APIs will go nowhere. Why would an independant Applicaiton company spend money implementing SuperDCOM+++, when the existing interfaces just work? Imagine, the Windows API could actually stabilize (and be reverse engineered)!
This would essentialy kill Windows, Inc. as a software business, BTW. The "Stable API" model only works for UNIX because the companies don't give a crap about the commodity market and are chasing the very profitable high-end hardware market. The Unix APIs themselves have very little commercial value as software (see Linux).
Windows Inc. -- Win9x, WinNT, WinCE, IE (at this point it is in the OS), IIS, VisualStudio dev tools, BackOffice server apps, Outlook (Exchange client).
(Note that you could even imagine a "BackOffice, Inc." that included BackOffice, Outlook, and VisualStudio.)
Office Inc. -- MS Office, all other business and consumer applications, MSN, WebTV, input devices.
Other Microsoft has been making large investments in cable and telco companies, as well has hollywood 'content providers'. In my opinion, these the spoils of monopoly and none of the new companies should be able to leverage interest here. Just as with the telco breakup, they should be required to disinvest and be prohibited from entering this market for a number of years.
I would think that breaking Microsoft up into "Windows, Inc." and "Office, Inc." would be a good thing from Bill Gates' standpoint as well. In the long term, the combined stock value of two market leading companies is probably greater than Microsoft alone.
Also, as you point out, they could focus on some of the decent server technology they have, and stop trying to produce their mutant network-centric "DNA" plan where everything on the network is tied insercurely and fragilely together at the Windows/Office client.
I would still prefer two branched distributions. I'd rather see more server-side stuff on a Server Distribution CD..., and more client/workstation packages (where to begin) included on the Workstation Distribution CD.
'Customized' distributions are an enormous advantage that Linux has over Microsoft, which basically has a one-size-fits-all product strategy (the only difference between NTS and NTW is the connection limit in NTW - the NT Server even comes out of box with workstation defaults, such as running foreground applications with a higher priority).
In a few years, I can imagine a "Linux Firewall" distribution, a Linux "Secure Mail Server" distribution, an ecommerce distribution, any number of customized workstation installations, and so on. No need to have a guru on hand, just drop in the CD and answer a few questions and you are running with a reasonable assurance that you don't have a gopher server in the background or 50 MB of extra applications you don't need. Sure this is fragmentation, but the lower cost of ownership a tailored distribution would bring makes up for it. Plus it would still be the same Linux underneath.
And for those who need more complex or customized setups, or just want the kitchen sink, the classic distributions (RedHat, Suse, etc) aren't going anywhere.
A better analogy might be the 4 door compact sport utility vehicle. (AMC, mid-80s.) It's been ripped off by just about everyone, but Jeep still has a pretty big market share.
Well in a pretty short period of time, the average PC monitor has gone from about 60 dpi to 100 dpi+. Eventually we're going to have display technology which gives us "laser print" resolution (200-300 dpi). Maybe then it will become obvious to the overlords of the X Window System (the commercial UNIX companies) that X's font handling and general pixel-dependancy is broken.
There was a ThinkPad with a removable cover some years ago too. You could pop it off and then 'invert' the transparent screen for use with an overhead projector. Kinda cool, but probably not that practical in real life because of all the rigging necessary to hang the thing off the overhead.
and the FCC who was supposed to encourage competition is naively allowing these mergers
Don't fool yourself. The Communications Act of 1996 was written with millions of dollars of corporate payola. "Competition" actually means massive consolidation.
Yeah, it's easy. So easy that RedHat should update their installer so that it automatically does this. Unlike NT, the infrastructure is already in place. (I don't see too many IIS hotfixes on Windows Update!)
Security holes are a fact of life, and 'critical' patches are a routine event. It's time that all OS vendors start treating them as such, as opposed to the the current strategy of putting the burden on each and every administrator out there.
In a couple years, Compaq will be supporting SCO on IA32, Monterey on IA64, Tru64 on Alpha, VMS on Alpha, probably Linux on all three, and NT on both IA32/IA64. That's lots of products that function in the same market space. The bottom line is that something has gone horribly wrong with the Compaq DEC merger, from a strategy perspective.
The theory was that Compaq-DEC would be able to compete with IBM. While supporting lots of OSes, IBM at least has a coherent product lineup from top to bottom. Compaq probably just has confused salespeople.
Slashdot Mirror
The Linux source might be GPL but at the end of the day if Linus or his advisors don't like the changes you suggest it won't get into the official kernel.
True, but many Linux vendors don't ship an 'official' kernel and instead fork it with add-ons that Linus hasn't gotten around to including yet. (e.g. ISDN and NFS fixes)
One thing I have noticed reading linux-kernel is that the development process is only 'open' to those coming with working code. That is, bug reports and feature requests are usually ignored unless there is a patch attached to the message. (Fine, because linux-kernel is about engineering, not marketing.) Commercial vendors, on the other hand, have mechinisms in place for bug reports and 'enhancement requests'. However, such reports typically disappear into a black hole from a user perspective - there is very little feedback until a fixlist emerges for the new version.
Sun is embarking on an opportunity take the advantages of both open source and commercial development practices. However, if Solaris patches disappear into the same black hole as enhancement requests, it's not going to work.
I would hope that (1) Sun starts an open list, similar to linux-kernel, to discuss Solaris development issues and review patches, and (2) there is an open central repository where users can download submitted patches. (Similar to the way Linux users get access to 'unapproved' stuff like ISDN or NFS.) Even if someone can't redistribute "MacMillian Solaris", if users have to wait for Sun's blessing to get access to patches, it's not going to work.
I have my doubts that Winblows95/98 is an example of good software engineering, its more like a HACK job, or would that be better said as ONE_BIG_KLUDGE? ;-)
Don't forget that the #1 design goal of Windows 95 was to be backwards compatible with 10 years of hardware drivers and DOS/Win16 programs. They got that part right.
The #2 goal was to provide a similar API to Windows NT so that at some point in the future they could obsolete DOS/Windows. They got that mostly right, although 9x makes so much money that DOSWin isn't extinct yet.
Later on, with DirectX, they wanted to provide a better game environment than DOS could provide. I haven't seen too many DOS games lately, so I assume that's going well.
Note that security or reliablity were never high up on the design goal list for Win 9x. They already were selling Windows NT, you know, but even to this day people choose backwards compatiblity and games over a superior OS.
If this is the case then sun should stop deceiving people into thinking that this is open source. Whenever Sun talks to the press ...
/. a few days ago) makes the comparision with Linux, but they're ZDNet, not Sun. They quote someone from Sun:
You are going to have to provide a reference to prove that - I couldn't find one.
Sun's standard PR bit on SCSL says:
The CSL model is the result of Sun's continued commitment to the open development of key technologies, and is based on Sun's community source principles, which include immediate open access, increased innovation, faster commercialization, and access for students.
ZDNet (referenced by
At the end of the day, the issue is how you allow innovation, but also have a reasonable process by which the community sticks together on the core as it evolves," said Anil Gadre, general manager of Sun's Solaris division.
Neither statement looks like an appeal for the support from Open Source (TM) advocates or developers. AFAIK, they didn't enlist Eric Raymond or make any appeals to Linux developers.
And, they will probably get some free labor, although it will be from commercial developers and hardware manufacturers, not the OSS crowd. Which makes sense, because by-in-large Sun doesn't make any money from the OSS crowd, so why should they care?
While Sun is allowed to derive benefit from the community, the community is placed in a state of perpetual legal risk
Did it occur to you that perhaps the "community" referred to in the Community Source Licence is not the same community that comprises the Linux/GNU user base? Just because they mention "community" and "source", there is no reason to make the egotistical leap that they are talking about the "Open Source Community".
There is a very large, and very valid community that can benefit from Sun's SCSL. Namely people who depend on Sun products and don't give a shit about redistribution or forking or winning the hearts and minds of Linux users. If that's not you, why should you care? Source for Solaris can mean bugfixes today and not when Sun gets around to it, allows more code optimization, makes it easier to write device drivers, and so on.
(Note - Same argument goes for those who flame Apple's source licence.)
Furthermore, you appear to be demanding that Sun should start giving away commercial software simply because it would be useful to you and others. Why pick on Sun's SCSL chart module? Why not demand the dozens of other commercial closed source chart programs (like MS Excel)? Are you trying to flame Sun until they GPL all of their software just because the SCSL makes them seem amenable?
You might be right. There are already NT4 "Post-SP6" hotfixes out, and SP6 hasn't even been released yet.
Isn't this argument over software subwoofer detection rather lame? After all, this problem was solved in a "plug and play" manner in hardware 40 years ago. (Or was someone too cheap to solder in a $.10headphone jack?)
It's kind of interesting that none of the commercial vendors have ponied up the money for a TPC benchmark on Linux. I would think that it's a given that Oracle/Linux beats Oracle/NT and probably MSSQL/NT as well....
Perhaps RedHat (et al) are too busy chasing the "workgroup server" and small business markets to bother with trying to sell database servers.
Don't forget #8 - Any new feature you want will be in the next version, which is due in six months from any given point in time.
I recall 1994, when a NT administrator could demand 25-50% more money than a NetWare administrator. As NT's usage spread, and Microsoft was successful in promoting it's education programs, the price of NT administrators dropped through the floor.
UNIX (and AS/400, VMS, etc) admins are expensive because getting admin access to a Unix system is difficult. so fewer people learn. A vicious catch-22.
Linux, on the other hand doesn't have the access problem. I'm sure that in a couple years there wil be enough Linux admins that labor costs will be comparable to other PC NOSes.
2) Remove all private interfaces between the operating system development group and all other software development groups at Microsoft
You are hitting at the heart of the integrated product strategy at Microsoft. Consider the introduction of OLE2 in 1993. Microsoft released the operating system feature (OLE) as part of MS Office, and as a tool in VisualBasic. Users instantly had an example of the functionality and could go out and implement it themselves.
Compare this to Apple OpenDoc. Apple released the API, but refused to 'integrate' it in ClarisWorks or the MacOS Finder. Result was that nobody really understood the possibilities and very few applications ever bothered with the API.
If a brick wall is instituted between the MS OS group and the Apps group, the result is that most new OS APIs will go nowhere. Why would an independant Applicaiton company spend money implementing SuperDCOM+++, when the existing interfaces just work? Imagine, the Windows API could actually stabilize (and be reverse engineered)!
This would essentialy kill Windows, Inc. as a software business, BTW. The "Stable API" model only works for UNIX because the companies don't give a crap about the commodity market and are chasing the very profitable high-end hardware market. The Unix APIs themselves have very little commercial value as software (see Linux).
If MS were broken up into two chunks:
Windows Inc. -- Win9x, WinNT, WinCE, IE (at this point it is in the OS), IIS, VisualStudio dev tools, BackOffice server apps, Outlook (Exchange client).
(Note that you could even imagine a "BackOffice, Inc." that included BackOffice, Outlook, and VisualStudio.)
Office Inc. -- MS Office, all other business and consumer applications, MSN, WebTV, input devices.
Other Microsoft has been making large investments in cable and telco companies, as well has hollywood 'content providers'. In my opinion, these the spoils of monopoly and none of the new companies should be able to leverage interest here. Just as with the telco breakup, they should be required to disinvest and be prohibited from entering this market for a number of years.
I would think that breaking Microsoft up into "Windows, Inc." and "Office, Inc." would be a good thing from Bill Gates' standpoint as well. In the long term, the combined stock value of two market leading companies is probably greater than Microsoft alone.
Also, as you point out, they could focus on some of the decent server technology they have, and stop trying to produce their mutant network-centric "DNA" plan where everything on the network is tied insercurely and fragilely together at the Windows/Office client.
I really doubt it. Except for www.lotus.com and small parts of other sites, Lotus Domino is used almost exclusively for intranet applications.
(Which means that content regulation doesn't affect Lotus' market one way or another.)
I would still prefer two branched distributions. I'd rather see more server-side stuff on a Server Distribution CD ..., and more client/workstation packages (where to begin) included on the Workstation Distribution CD.
'Customized' distributions are an enormous advantage that Linux has over Microsoft, which basically has a one-size-fits-all product strategy (the only difference between NTS and NTW is the connection limit in NTW - the NT Server even comes out of box with workstation defaults, such as running foreground applications with a higher priority).
In a few years, I can imagine a "Linux Firewall" distribution, a Linux "Secure Mail Server" distribution, an ecommerce distribution, any number of customized workstation installations, and so on. No need to have a guru on hand, just drop in the CD and answer a few questions and you are running with a reasonable assurance that you don't have a gopher server in the background or 50 MB of extra applications you don't need. Sure this is fragmentation, but the lower cost of ownership a tailored distribution would bring makes up for it. Plus it would still be the same Linux underneath.
And for those who need more complex or customized setups, or just want the kitchen sink, the classic distributions (RedHat, Suse, etc) aren't going anywhere.
A better analogy might be the 4 door compact sport utility vehicle. (AMC, mid-80s.) It's been ripped off by just about everyone, but Jeep still has a pretty big market share.
perhaps we should instead be looking for printer drivers that accept something like HTML, and render it as they see fit for the target device.
Like, uhh, PostScript?
Well in a pretty short period of time, the average PC monitor has gone from about 60 dpi to 100 dpi+. Eventually we're going to have display technology which gives us "laser print" resolution (200-300 dpi). Maybe then it will become obvious to the overlords of the X Window System (the commercial UNIX companies) that X's font handling and general pixel-dependancy is broken.
The early Xerox networks ran XNS (Xerox Network Service?)
Which was the basis for the Novell IPX protocol, just to add to the things that Xerox never made any money from.
There was a ThinkPad with a removable cover some years ago too. You could pop it off and then 'invert' the transparent screen for use with an overhead projector. Kinda cool, but probably not that practical in real life because of all the rigging necessary to hang the thing off the overhead.
(Model number was 755AV or something.)
Of course, Windows green is people.
There used to be a public microsoft.com host running Linux. It was a standard chestnut on
(I think it was a news server or something. Nothing that exciting.)
and the FCC who was supposed to encourage competition is naively allowing these mergers
Don't fool yourself. The Communications Act of 1996 was written with millions of dollars of corporate payola. "Competition" actually means massive consolidation.
Yeah, I've looked around MS's site, and the only central location which lists current hotfixes that I've found is the FT P directory!
Yeah, it's easy. So easy that RedHat should update their installer so that it automatically does this. Unlike NT, the infrastructure is already in place. (I don't see too many IIS hotfixes on Windows Update!)
Security holes are a fact of life, and 'critical' patches are a routine event. It's time that all OS vendors start treating them as such, as opposed to the the current strategy of putting the burden on each and every administrator out there.
In a couple years, Compaq will be supporting SCO on IA32, Monterey on IA64, Tru64 on Alpha, VMS on Alpha, probably Linux on all three, and NT on both IA32/IA64. That's lots of products that function in the same market space. The bottom line is that something has gone horribly wrong with the Compaq DEC merger, from a strategy perspective.
The theory was that Compaq-DEC would be able to compete with IBM. While supporting lots of OSes, IBM at least has a coherent product lineup from top to bottom. Compaq probably just has confused salespeople.