People: Microsoft _invented_ the term "server OS" to cover up the single-user-ness of their flagship product.
Now that's ridiculous. The "Server OS" term came into being long before Microsoft had a product in that space, before Microsoft even included networking with their client products. Does anyone remember Netware? Banyan? Actual server operating systems, since one couldn't actually use the server as a personal workstation at the same time.
By the way, my original comment was meant as a humorous riposte to your obvious off-topic advertising of a closed, commercial product on a thread about Open Source Software and its future. I posted it with my tongue firmly planted in my cheek. The fact you couldn't read that into the post, shows that maybe you're spending too much time around the marketing and VC guys and not enough time coding, Quaking and enjoying life with the rest of us.
The size and foaming rabidity of your reply is clear evidence of your "good humor". And evidence of the trolling nature of your initial message. 'Course, you probably don't see it that way. Your religious fervor doesn't admit of the possibility for even a fraction of a second you might have launched an attack in error and at the wrong target.
And let's not even talk about the ad hominem attacks.
The sole focus of my reply, as you are well aware, was and is contributions to gcc, binutils, and egcs. Open software, all. Covered by GPL, even. The source for which resides on the above-referenced web site, in accordance with the GPL. (And if someone asks nice and sends me a check to cover the cost of a CD-W blank and my time, I'll burn it and mail it.)
Yes, the context in which the changes were made is in support of what is, to some extent, a closed source product. The general applicability of the changes, as evidenced by ongoing work on the egcs mailing list, is far wider than that one closed-source product.
Had anyone thought for even a moment that my messages were ads, I'm quite certain they would have been moderated into oblivion. Check the ratings.
By the way, do you spread that venom and bile to other "closed" systems, like Solaris? After all, they had the temerity to make it possible to run Linux binaries (which is, in your world view, The One Truely Politically Correct format)? Or Caldera? (Don't they have proprietary, closed-source programs as part of their package?) TurboLinux (same)?
Or is it reserved just for anything that runs on the Evil OS From Redmond? And if that's the case, do you heap scorn upon the fine folks from Cygnus, whose cywgin32 environment runs on said OS and who sell (gasp!) a non-libre license for said environment?
There are a few reasons why one might want an Interix-native version of emacs to run on NT.
Win23 Emacs contains a whole mess of compromises; places it has to expose the "win32-nature" of the environment. Drive letters, line termination characters, that sort of thing. Because Interix provides a full UNIX-like environment, there are no compromises in behavior; it looks and feels exactly like a UNIX version of emacs, because it is.
Emacs is more than an editor; between shell modes, and debugger modes, and mail modes, and program compilation and build modes, the whole environment assumes there's a real UNIX system underneath the covers. With Interix, there is.
The emacs source code is riddled with #ifdefs to cope with the Win32 port. Bugfixes and enhancements to the mainline of emacs don't always make their way into those #ifdefs; the functionality of the win32 port of emacs tends to lag behind that of emacs for UNIX systems. Because Interix provides a programmatic environment just like a UNIX system, the #ifdefs required to adapt emacs to Interix are, in theory, substantially smaller than those required for the Win32 port; in fact, from a source code portability standpoint, Interix looks like "just another UNIX platform" instead of "whoa, that's a completely different world" like Win32.
Until, a year later, someone wants not just a feature addon, but a larger change that will require some architectural changes in the software. Now what? I'm off to the Bahamas and can't be reached, and no one else really has any real insight into what goes on in my code.
You're kidding, right? Cygnus makes a very healthy living supporting OSS; stuff they've written, stuff other people have written. If a program is important (e.g. egcs) to some company's business, they'll pay someone, anyone, to provide it.
How do we ensure, that a program developed thru these new efforts gets maintained, or at least, keeps it's maintainers?
Use cosource/sourceXchange/any-other-means to tell the same community "I'm willing to pay someone to support this code". Or develop your own expertise and support it yourself.
Compare this to the proprietary source model. The vendor assumes that, if you want the product, you also want some low level of support; the cost of that low level of support gets into the price of the product. The vendor then sets a higher level of support and charges another fee to obtain that level, adjusting the level and price until profit is maximized.
In the cosource/sharedXchange model, there are no assumptions. Desire for and willingness to provide support are made explicit; an efficient market in "what does support mean to you?" evolves. I can see some people bidding $x each for "fixes to these ten bugs"; I can see other people bidding $10x for "fix any bug I can reproduce for you within 5 working days".
None of the above is profound; it's been said before, and amply demonstrated, by Red Hat, and Cygnus, and a whole lot of other people and companies.
A quick search of the above-cited URL for the keyword "bounty" yields this page.
We paid off on egcs, for sure. We're now in the process of working with the egcs gang to roll those changes (and a bunch more) into the 1.2 egcs release (well, now the gcc 2.95 release). Our changes do more than just support Interix; they fix a whole lot of broken stuff for support of Alpha processors in 32-bit mode (as used by Cygwin32 as well as Interix), some floating point bugs, some g77 problems, etc. The "leg up" on egcs 1.1 we got from the port was worth more to us than the money we paid out; the amount of time it took the person doing the work was less than he believed the money justified. Everyone came out ahead.
The emacs port was close, but no cigar; not enough worked. I'm pretty sure we paid out on TeX.
No. It's something awfully close to Unix (but we're not allowed to apply that trademarked adjective to our product... yet).
It just happens to be another personality layered on top of the microkernel at the heart of Window NT.
I'm not going to justify Interix here. The website does a good job of it, and that's not what slashdot is for, anyway. If you've an open mind, read the web site. If you lack that open mind, go away and do not troll here.
Speaking as an executive of a company that has paid individual developers for exactly this purpose, I'm thrilled to see an infrastructure develop to make things easier.
Last year my company sponsored what we called a "Linux Bounty"; we identified a dozen or so Open Source applications that we wanted ported to our operating system. We put dollar amounts on each and set up some simple rules:
First acceptable implementation gets paid
We decide what's acceptable
We'll update in real-time a list of projects for which we have received implementations
Whatever copyright terms applied to the original source base applied to the submitted code
We got a half-dozen implementations; for those which were useable, we paid. The most expensive item cost me $1000 out of my budget, and was cheap at the price.
The biggest unanswered question about sourceXchange indeed lies in the matter of peer review. If I'm going to write a check for a few grand, I'm going to want to have some degree of trust in the technical judgement of the person or persons deciding if the submitted work meets the requirements. I expect I'm going to have to write wickedly precise requirements with respect to coding standards; things like "no compiler warnings beyond those generated by the starting source base", etc. Things that I need guaranteed if I'm going to be able to support the resulting software.
All in all, though, I expect to be a sponsor of a variety of projects. Start small, work my way up as I gain trust in the process. I don't like the idea of starting small; it will cost me time I can ill-afford. But it's the only way to mitigate the risks involved with someone else making an accept/reject decision on my behalf.
It also renders the need to actually run the Linux kernel as moot -- Linux binaries will run quite nicely on any old proprietary operating system that puts in the appropriate hooks.
That makes the assumption that Linux stands still, i.e. that the OS itself no longer advances. That's not a good bet. I expect to see Linux continue to grow; just look at the stuff that SGI is throwing into the pot (and it's not random crap; the filesystem is pretty killer).
I'm perfectly happy to see IBM and Sun modify their proprietary operating systems to add features developed elsewhere in an open process and demanded by users and their applications. It would be pretty silly of them to make those new features available only in Linux emulation mode; they'll have to integrate them with the rest of their OS. The quality (or lack thereof) would certainly factor into a decision about which OS to run; it might lead one to stick with Linux, which isn't burdened by all that proprietary (and historical) baggage.
The implication: control of the long-term direction of Unix is removed from the hands of a small group of vendors and placed into the hands of a group of smart people who are willing to build what they want and experiment with it until it's right. Sounds like a winner to me.
(Lest you think I'm ignoring the formal standards aspect of What Is Unix - those folks, the IEEE, The Open Group, have pretty much recognized that one can only standardize real, honest-to-god, deployed and existing practice. Linux is the best way of deploying new technology and seeing if it works; based on that real-world feedback, better standards can (and will) be written.)
But it is clear that the people that the IEEE does serve are their members only. That's not including me or thousands of other/. readers.
You know that ethernet that you probably use, that/. uses, that everyone else uses? IEEE Std. 802.3. The standard for UNIX? IEEE 1003. SCSI? That's an IEEE standard, too. Standards development isn't cheap; editors cost money. (They gotta eat, too.)
Standards development is just one of the services the IEEE performs for the industry at large. There are others.
Full disclosure: I chair an IEEE standards working group. I personally know people who work for the IEEE standards department. It's likely that little money from the IEEE pubs department goes towards supporting the standards effort at the IEEE; this is merely an example.
those banner ads at the top of these pages probably go to fattening the pockets of the "team".
Actually, the income from banner ads probably goes to fatten the pockets of the ISP where/. colocates its server. And it probably paid for the server. And the ever-increasing disk space required to archive insightful (and less so) comments.
Slashdot Mirror
Now that's ridiculous. The "Server OS" term came into being long before Microsoft had a product in that space, before Microsoft even included networking with their client products. Does anyone remember Netware? Banyan? Actual server operating systems, since one couldn't actually use the server as a personal workstation at the same time.
The size and foaming rabidity of your reply is clear evidence of your "good humor". And evidence of the trolling nature of your initial message. 'Course, you probably don't see it that way. Your religious fervor doesn't admit of the possibility for even a fraction of a second you might have launched an attack in error and at the wrong target.
And let's not even talk about the ad hominem attacks.
The sole focus of my reply, as you are well aware, was and is contributions to gcc, binutils, and egcs. Open software, all. Covered by GPL, even. The source for which resides on the above-referenced web site, in accordance with the GPL. (And if someone asks nice and sends me a check to cover the cost of a CD-W blank and my time, I'll burn it and mail it.)
Yes, the context in which the changes were made is in support of what is, to some extent, a closed source product. The general applicability of the changes, as evidenced by ongoing work on the egcs mailing list, is far wider than that one closed-source product.
Had anyone thought for even a moment that my messages were ads, I'm quite certain they would have been moderated into oblivion. Check the ratings.
By the way, do you spread that venom and bile to other "closed" systems, like Solaris? After all, they had the temerity to make it possible to run Linux binaries (which is, in your world view, The One Truely Politically Correct format)? Or Caldera? (Don't they have proprietary, closed-source programs as part of their package?) TurboLinux (same)?
Or is it reserved just for anything that runs on the Evil OS From Redmond? And if that's the case, do you heap scorn upon the fine folks from Cygnus, whose cywgin32 environment runs on said OS and who sell (gasp!) a non-libre license for said environment?
- Jason
You're kidding, right? Cygnus makes a very healthy living supporting OSS; stuff they've written, stuff other people have written. If a program is important (e.g. egcs) to some company's business, they'll pay someone, anyone, to provide it.
Use cosource/sourceXchange/any-other-means to tell the same community "I'm willing to pay someone to support this code". Or develop your own expertise and support it yourself.
Compare this to the proprietary source model. The vendor assumes that, if you want the product, you also want some low level of support; the cost of that low level of support gets into the price of the product. The vendor then sets a higher level of support and charges another fee to obtain that level, adjusting the level and price until profit is maximized.
In the cosource/sharedXchange model, there are no assumptions. Desire for and willingness to provide support are made explicit; an efficient market in "what does support mean to you?" evolves. I can see some people bidding $x each for "fixes to these ten bugs"; I can see other people bidding $10x for "fix any bug I can reproduce for you within 5 working days".
None of the above is profound; it's been said before, and amply demonstrated, by Red Hat, and Cygnus, and a whole lot of other people and companies.
We paid off on egcs, for sure. We're now in the process of working with the egcs gang to roll those changes (and a bunch more) into the 1.2 egcs release (well, now the gcc 2.95 release). Our changes do more than just support Interix; they fix a whole lot of broken stuff for support of Alpha processors in 32-bit mode (as used by Cygwin32 as well as Interix), some floating point bugs, some g77 problems, etc. The "leg up" on egcs 1.1 we got from the port was worth more to us than the money we paid out; the amount of time it took the person doing the work was less than he believed the money justified. Everyone came out ahead.
The emacs port was close, but no cigar; not enough worked. I'm pretty sure we paid out on TeX.
No. It's something awfully close to Unix (but we're not allowed to apply that trademarked adjective to our product... yet).
It just happens to be another personality layered on top of the microkernel at the heart of Window NT.
I'm not going to justify Interix here. The website does a good job of it, and that's not what slashdot is for, anyway. If you've an open mind, read the web site. If you lack that open mind, go away and do not troll here.
Last year my company sponsored what we called a "Linux Bounty"; we identified a dozen or so Open Source applications that we wanted ported to our operating system. We put dollar amounts on each and set up some simple rules:
- First acceptable implementation gets paid
- We decide what's acceptable
- We'll update in real-time a list of projects for which we have received implementations
- Whatever copyright terms applied to the original source base applied to the submitted code
We got a half-dozen implementations; for those which were useable, we paid. The most expensive item cost me $1000 out of my budget, and was cheap at the price.The biggest unanswered question about sourceXchange indeed lies in the matter of peer review. If I'm going to write a check for a few grand, I'm going to want to have some degree of trust in the technical judgement of the person or persons deciding if the submitted work meets the requirements. I expect I'm going to have to write wickedly precise requirements with respect to coding standards; things like "no compiler warnings beyond those generated by the starting source base", etc. Things that I need guaranteed if I'm going to be able to support the resulting software.
All in all, though, I expect to be a sponsor of a variety of projects. Start small, work my way up as I gain trust in the process. I don't like the idea of starting small; it will cost me time I can ill-afford. But it's the only way to mitigate the risks involved with someone else making an accept/reject decision on my behalf.
Jason
That makes the assumption that Linux stands still, i.e. that the OS itself no longer advances. That's not a good bet. I expect to see Linux continue to grow; just look at the stuff that SGI is throwing into the pot (and it's not random crap; the filesystem is pretty killer).
I'm perfectly happy to see IBM and Sun modify their proprietary operating systems to add features developed elsewhere in an open process and demanded by users and their applications. It would be pretty silly of them to make those new features available only in Linux emulation mode; they'll have to integrate them with the rest of their OS. The quality (or lack thereof) would certainly factor into a decision about which OS to run; it might lead one to stick with Linux, which isn't burdened by all that proprietary (and historical) baggage.
The implication: control of the long-term direction of Unix is removed from the hands of a small group of vendors and placed into the hands of a group of smart people who are willing to build what they want and experiment with it until it's right. Sounds like a winner to me.
(Lest you think I'm ignoring the formal standards aspect of What Is Unix - those folks, the IEEE, The Open Group, have pretty much recognized that one can only standardize real, honest-to-god, deployed and existing practice. Linux is the best way of deploying new technology and seeing if it works; based on that real-world feedback, better standards can (and will) be written.)
Standards development is just one of the services the IEEE performs for the industry at large. There are others.
Full disclosure: I chair an IEEE standards working group. I personally know people who work for the IEEE standards department. It's likely that little money from the IEEE pubs department goes towards supporting the standards effort at the IEEE; this is merely an example.