I guess, maybe I obscured my point. Linux is an OS that was developed initally to run on a 386SX...or, extremely low end system. UNIX in general can quite easly blur the line between workstation and server. I can run a 386SX20 w/ 6M RAM, put apache on it, imapd, and basic server functions like shells and call it a server. And, in truth it IS a server, which can scale fairly well up to 10 users or so for pop3, and do a pretty decent job for web serving (easily taking about 20 hits per minute, provided it's static pages, and not cgi or database stuff). As a matter of fact, I have done just that in the past, and it worked well. And the hardware costs were under $70 USD. And there is no way in the world I would ever call that box something that could be used as a "WorkStation."
Battering about the term "Server" to justify the high price tags of new brand name (DEC, Compaq, IBM, Dell, etc) Linux boxes doesn't hold well in my book. I know people who keep thier workstations load peged for weeks at a time doing calculations, rendering, etc. And I see servers that couldn't handle workstations loads that easily handle massive server loads (Alpha box in ND-HECN that has well over 1000 users, but my K6 will outrun for pure CPU power, the CDROM.COM FTP server that holds records for transfer in a day that is a dual PPro, and wouldn't hang for massive rendering or number crunching quite as well as some workstations sitting around my department).
The primary diffrance that "they" are suggesting makes a server a server is RAID, multiple network interfaces, etc... But, at the same time, not all of thier "server" offerings have these features standard (What makes Dell's Linux server a server anyway?).
I think they don't want to do Linux workstations (IMHO) because they fear the "My Big Gulp won't fit in my new 20X Cup Holder" phone calls. Slashdot is a Dual Pentium II 233 for crips sake, and it's a "server," and no one would argue that it's really just a workstation, and not a server.
I guess, in my mind, no one has doubted Linux's ability to be a server for a couple years now, but it seems that's where it ends. And, to me, that's not news. Saving a few thousand dollars on a high end server by choosing Linux insted of DEC-UNIX or AIX on your server may be big news to some people. But it's not big news to the general population, because, more people use PC's/Home Office/Workstations that actually run thier own server. And, today, Linux is make BIG inroads in the workstation market, and that's where it's truely starting to shine. It has always shined as a server. But now that it's shining in the scientific labs doing numbercrunching, theoretical predictions, and data analysis as a WorkStation. Now it's shining in CAD/Rendering/Media arenas as a WorkStation. That's a bigger market. That's bigger news. That's where I would like to see some of the big "brand names" take notice and become active in support.
And, as far as that goes, a "entry level" server would be great to see too. I guess this is happening with Corel's Netwinder, and the Qube... But, I think buisnesses that have 5-20 people in the office could benifit soo much more substantually from Linux than they realize. If there were a commercial solution with ONE server running 10-20 Xterminals (like Mechanical Engineering Dept. at the University of Minnesota has done), That would turn heads, and make news, and prove that the idea of "workstation" vs. "server" is a very blurry line, and you have to think more in terms of "networked solutions" than "My Workstation" and "our company's server." Sometimes the server only needs to be a little 486, and the workstations need to be quad Xeons... but sometimes the workstations only need to be 486's, and the server needs to be the quad Xeons or RS/6000 systems. How can anyone draw such lines as "workstation" and "server" and use that to defend what market and price range they are focusing on?
I have been looking at the little 43P for a while, and I assumed that AIX was part of the reason that a box with these specs cost so damn much. So, thinking something like this running Linux would really cut down on costs, I had hopes someday I might get one. Now after reading that linked report, I wonder. What systems were they talking about exactly that are $15,000! Can't be the little ones.
I do like the fact that Compaq, Dell, IBM, SGI, HP, Apple, others now offically support Linux in some way. But the way makes me wonder. Dell won't do a kernel for duals, Apple doesn't make getting a "MkLinux" box easy, it's a "do it yourself" kind of thing. The IBM report linked seems to be WAY on the high end for a typical user (Heck, even the box that runs SlashDot can't be worth $15,000, is it?).
My hopes now are with companies like VAResearch. Hopefully someone will come along soon that will sell boxes that truely span the $700 to $2000 range, that come with Linux professionally configured, with good support, and offer some hardware choices like Motorola, AMD, Cyrix, Alpha, MIPS, etc... But, as the "Big Names" come into the Linux ring, I thought that would happen, and in truth, it's getting farther from the truth it seems...
Currently, our department (University of North Dakota Chemistry) is considering buying a dual Xeon system for number crunching. There is a big debate about VAResearch vs. Dell (Dell has a deal with our school, University of North Dakota). Larry M. Augustin (of VAResearch) is willing to work with us, if the people here in charge of the decision would only get off ther butts and call him. Dell has shown no interest in supporting something like we want, but now makes this announcement.
Looking at $6,000 to $12,000 boxes all over the place, wondering what commercial Fortran compilers will do for us, what will be better, NT, Linux, Single Xeon 2M cache, dual 1M Xeon systems, all kinds of details that are very specific to the data arrays of the calculations, and leaving many unanswered questions. Double the price, double the preformance? Maybe, maybe not, maybe better compilers are the answer, maybe not, maybe we should just get another SGI. It's a total mess of a decision here. And, right now, since the real "cheifs" (pointy hair bosses) are making the decisions of who to even gets to talk, it get's even tougher.
I tell you, Dell not supporting more graphic cards, and not supporting Dual will play heavy against them in our decision. At the same time, "loyalty" plays it's ugly hand in favor of Dell for no good reason.
Some times, even when it looks like good news at first glance, it only makes the situation worse. The "Pointy Hair Boss" syndrome seems to play more heavy into these decisions than anything Dell or VAResearch can do/say/support. It's all well and good as far as Linux PR goes. But when you need the box NOW, and avaliability of Dell's limited support is coming "in a few weeks," and you can't even get the pointy haired bosses to TALK to the suppliers, this news just makes me wanna scream.
Sorry, I think I am babbleing incoherently now, I'll shut up, I just had to vent. Larry, if your out there, I am still fighting for you guys here, I hope I get them to pick up the damn phone. (Damn the phone, which BTW, Dell isn't answering today when they called to get details about the Dual's and Linux. Strike TWO for Dell.)
You may want to Check thier Pricing/Configuration before you say it's out of line. I see a heck of a lot of hardware there for under 1/2 of the $50,000 you mention.
For example, in December VA began shipping the VArServer 4100 system, a Red Hat Linux server that features four 450-MHz Intel Pentium II Xeon processors with 1GB of RAM on each, upgradable to 2GB. Pricing starts at $49,995. 1GB of RAM _ON_ each? What? I do believe they meant to say 1M full speed Cache Ram on each. If they stuck 1G on a CPU, they would be like world record holders or something....
I disagree, see my previous post one level up, same level as yours. I know my way around a box, and I can spark a shell and insert a kernel module on my own during install. But when they are rejected, the whole thing is fuxered.... Not just some little package that needs to be updated.
I got very anoyed with RedHat 5.2 this weekend. ON THREE systems with THREE diffrent ethernet cards, I grab the boot disk, go to do an nfs install (on two of the boxes) or ftp install (on the third) and the bootdisk/supplemental disk won't load the kernel modules for any of the three ethernet cards I wanted to use (NE2000 ISA, NE2000 PCI, and WD8013)!
So, I just grabed an old SuSE 5.1 boot disk, and it worked flawlessly (and on ONE disk, yes, a ONE FLOPPY, NO CDROM complete install). Of course, then I had to upgrade a ton of packages.
I have a great deal of respect for Red Hat, and I wish them and Bob Young the best of luck. But, putting out 3-5 versions a year should NOT be a priority, getting it right should. 4.2 was good. Ever since that, I feel they have released to frequently, and too soon. IMHO 6.0 should have Gnome 1.0, and Kernel 2.2.x, and should be blanket tested in a widely distributed prerelease version before shipping. But, I guess I can't complain too much, 80% of my boxes run Red Hat:-/
I don't fear Red Hat becoming an "Evil Empire at all. It's almost impossable. I would LOVE to see IBM sell a distribution (especially on thier G3 boxen!). I don't think that any of them can be a treat when the source is GPL'ed. The problem comes for the lack of a standard, and that is currently being addressed by the LSB committee (still working slowly and quietly in the background).
The only real threat would be that with the lack of something like the LSB, if some distribution grows too big, they BECOME the standard, (for example, you _know_ people would port to IBM's or Microsoft's version of Linux), and what if the big name distribution is non standard, leaving apps only working on that distribution.
dearth of applications that can exploit new SMP capabilities.
It's under the Linux Negitive side at the bottom of the page. Is there some exploit that screws up SMP Linux boxes running 2.2.0 or something? Or, is this his flowery way of saying that there aren't many things that really use SMP yet?
Believing it's the latter, isn't that totall BS because
Since in general source is avaliable for almost every app, and compile time flags, and smp optimization can be done for just about any app? (Doesn't SAL have a pretty big section about compilers for specific needs like this?
Even applications that are not SMP optimized benifit from good SMP implementation, since, this is suppose to be a server, and more than one job will be happening at a time, thus, there is more than one CPU to use?
Yes, not my subject, I could be totally wrong, but, I didn't think that that statement was totally true.
Thanks for the info, I now understand why I can set the read level to 1 insted of 0. Heh... Yea, sure, history.... Learn... Uh, and your IQ and educational level is? Funny? it's NOT, it's stupid. Chris Farly was famous for playing a moron and was funnier.
You do realize that doing ALL that html on a subject of cool cases, when wooden cases do exist, is idiotic, when the only thing that even HINTS at a joke is the URL? Man, if you think it's funny, I have a joke for you... "Jefferson and Hamlton walked into a pub, Jefferson said free the slaves, and no one said a word.... They were all in the army..." (if you don't get it, read some history).
If you put them into cases, why not just circumvent the bottlenecks, and just make the damn thing SMP?
I could be wrong, but I believe that it would be very difficult to get a SMP board that does over 4 CPU's, but using PC104's or SlotPC's, you can probably fit around 20 full systems in one case if you really tried.
100BaseT would be the bottleneck, your right, and you would want at least 100baseT and a good hub IN the case also. It's tight, but it's not a bad idea. I would LOVE to actually see one (love to build one even more, but I lack the financial resources to do such a thing).
Which SGI case can you think of that is huge enought, and avaliable cheap? I am thinking Personal Iris, they should be damn cheap, but do you think it's big enough?
Actually, I wasn't thinking about an SGI case, but I was thinking of a single case Beowulf cluster idea.
Take the biggest Full Tower case you can find, and gut out all the bracketing, and make custom brackets to put in as many PC104's as you can fit (like this one). Then, use something like a 16M solid state IDE drive for each of the PC104's to put like/boot and what was needed on, and NFS mount a drive from each of them somewhere else in the case. Then put in maybe ONE bigger system (like maybe a PII?) to run as controller/terminal control(running X), and also stick the hubs in there.
You might be able to squeze in like 10/20 full computers into one case with this idea, and use 100baseT to connect them, NFS to a few big SCSI drives in the case, and have it completely self contained! You probably would be limited to 300MHz socket 7 CPU's, but, w/ 10-20 of them in one case, working together the Beowulf way, it might totally rock.
This is the dumbest fake I have ever seen? When the html is soo anoying you cna't even read the text because of the layers, how are most people suppose to figure out it's a fake?
Wooden cases ARE made, by a few people, including IE NET, and I find them a fairly cool thing, in thier own way. So, I don't see the humor in spoofing something that is real. I mean, it's not a stupid idea, it's a cool idea. And it's out there.
Someone might have to explain what is suppose to be funny about this to me... Like, would it be funny if I put up a site that said I was building and selling MIPS servers, and did elaborate html, with photos and details, but I didn't really make or sell them? I don't think it's funny, I just don't get it. It's stupid.
I happen to be sitting at a AMD-K6-2-3D(333MHz) with 128M RAM, a 4M MilleniumII video card, and 15" Sony 100ES monitor. Now, normally, w/ gimp or something, I never have any problems, and the box handles things well. Little did I know that a IDIOTICLY designed web page would peg my load, and render this box completely useless. I couldn't even get to the damb "back" button in Netscape!
index.html was ok, I find it anoying to redirect to home.html, because it makes going "back" a tad more difficult, but then some crap blocked me from reading the home.html, so I clicked some button I didn't even know what was, and ended up in "products." Fine, it's a mess, but it looked like the product might be interesting. I started reading it, and all of a sudden I couldn't see anything, and my load was spiked, graphics were dumped all over the place in the browser window, slowly moving, and MY POINTER started "jumping" when I tried to get to the "back" button! I can run scripts in gimp, two or three at a time on several images, and have NEVER seen that happen!
This site should DEFINATELY be linked on a "what not to do" page about html design.
"and developing a server based on the Intel® IA-32 processor that will support Linux and Windows NT® operating systems."
So, what do you know about IA-32/Linux/SGI? What do you suspect the "server" design will be based on?
"As we enter the low-end of the market with 32-bit systems, we will support Windows NT and Linux, again giving our customers what they have been asking for," said Vrolyk.
I read comp.sys.sgi.hardware and the other comp.sys.sgi.* groups often... Do you? What do YOU see people asking for? SGI/LINUX servers? or SGI/Linux w/ X + GL?
Just because the web site isn't updated doesn't mean the project is dead. 3 months isn't that long to go without an update in HTML, and != no work on code.
PS. I think there may be an update by Monday, but I am not sure.
I did read the article. Yes it says servers. That's why the next obvious question is about X.
I have also read many other SGI comments and stories lately, and the burning question is X. Everyone is asking about X. So, lot's of people care about X. And when SGI announces that they will offically support linux for servers, people want to know "what about X?"
What's so hard to understand about people wanting X? Linux already runs on the Indy and the VPC, and therefore there is nothing new about being able to have a Linux/SGI server. The news story is about "SGI Offically Supporting Linux" and thier statement is that they will support it on SERVERS... Well, great, afaik, the support we recieve for IRIX is NOT great, and VERY expensive, so, although it's good PR for LINUX and SGI, it's not really that exciting of news when it comes down to the facts. What IS interesting about it is, with "offical" SGI support for linux in "some" ways, is this going to lead to what people REALLY want from SGI/LINUX... which is X.
What's so hard to understand about asking the _real_ question burning in peoples minds?
So, how far away are the SGI propriatry X drivers for the VPC now? Will GL be included? Assuming that SGI writes closed source drivers for X in Linux, is this a good or a bad thing? (good to have useable new hardware with superior graphics to existing x86 Linux, bad cause it might be closed?)
IMHO Educational institutions should do it the way UofM's Engineering Department did. With an equal budjet, you could server far more users and/or provide better computing power to all users. (considering probably 9 out of 10 workstations usually carry a light load at most times anyway..... but even still....)
Well, it's going to be a little complicated for some:-) I have actually "rpm'ed" in my last two kernel upgrades, which hasn't been that bad. But, going to 2.2.0 will go to IP chains (?) which might totally screw up my home LAN...
If it weren't for that, I would be happy if Red Hat did NOT come out with a new distribution just because of 2.2.0, and just put it in the updates. Then, they could push/wait for Gnome 1.0, and get rid of the awful default FVWM95 crud, and release "6.0" which would be cool. But... It's no doubt a quite complex issue, do you wait a little while and "get it all right" or do you jump on the bandwagon and try to be "The First" to do a 2.2.x distribution?
Slashdot Mirror
Battering about the term "Server" to justify the high price tags of new brand name (DEC, Compaq, IBM, Dell, etc) Linux boxes doesn't hold well in my book. I know people who keep thier workstations load peged for weeks at a time doing calculations, rendering, etc. And I see servers that couldn't handle workstations loads that easily handle massive server loads (Alpha box in ND-HECN that has well over 1000 users, but my K6 will outrun for pure CPU power, the CDROM.COM FTP server that holds records for transfer in a day that is a dual PPro, and wouldn't hang for massive rendering or number crunching quite as well as some workstations sitting around my department).
The primary diffrance that "they" are suggesting makes a server a server is RAID, multiple network interfaces, etc... But, at the same time, not all of thier "server" offerings have these features standard (What makes Dell's Linux server a server anyway?).
I think they don't want to do Linux workstations (IMHO) because they fear the "My Big Gulp won't fit in my new 20X Cup Holder" phone calls. Slashdot is a Dual Pentium II 233 for crips sake, and it's a "server," and no one would argue that it's really just a workstation, and not a server.
I guess, in my mind, no one has doubted Linux's ability to be a server for a couple years now, but it seems that's where it ends. And, to me, that's not news. Saving a few thousand dollars on a high end server by choosing Linux insted of DEC-UNIX or AIX on your server may be big news to some people. But it's not big news to the general population, because, more people use PC's/Home Office/Workstations that actually run thier own server. And, today, Linux is make BIG inroads in the workstation market, and that's where it's truely starting to shine. It has always shined as a server. But now that it's shining in the scientific labs doing numbercrunching, theoretical predictions, and data analysis as a WorkStation. Now it's shining in CAD/Rendering/Media arenas as a WorkStation. That's a bigger market. That's bigger news. That's where I would like to see some of the big "brand names" take notice and become active in support.
And, as far as that goes, a "entry level" server would be great to see too. I guess this is happening with Corel's Netwinder, and the Qube... But, I think buisnesses that have 5-20 people in the office could benifit soo much more substantually from Linux than they realize. If there were a commercial solution with ONE server running 10-20 Xterminals (like Mechanical Engineering Dept. at the University of Minnesota has done), That would turn heads, and make news, and prove that the idea of "workstation" vs. "server" is a very blurry line, and you have to think more in terms of "networked solutions" than "My Workstation" and "our company's server." Sometimes the server only needs to be a little 486, and the workstations need to be quad Xeons... but sometimes the workstations only need to be 486's, and the server needs to be the quad Xeons or RS/6000 systems. How can anyone draw such lines as "workstation" and "server" and use that to defend what market and price range they are focusing on?
I do like the fact that Compaq, Dell, IBM, SGI, HP, Apple, others now offically support Linux in some way. But the way makes me wonder. Dell won't do a kernel for duals, Apple doesn't make getting a "MkLinux" box easy, it's a "do it yourself" kind of thing. The IBM report linked seems to be WAY on the high end for a typical user (Heck, even the box that runs SlashDot can't be worth $15,000, is it?).
My hopes now are with companies like VAResearch. Hopefully someone will come along soon that will sell boxes that truely span the $700 to $2000 range, that come with Linux professionally configured, with good support, and offer some hardware choices like Motorola, AMD, Cyrix, Alpha, MIPS, etc... But, as the "Big Names" come into the Linux ring, I thought that would happen, and in truth, it's getting farther from the truth it seems...
Looking at $6,000 to $12,000 boxes all over the place, wondering what commercial Fortran compilers will do for us, what will be better, NT, Linux, Single Xeon 2M cache, dual 1M Xeon systems, all kinds of details that are very specific to the data arrays of the calculations, and leaving many unanswered questions. Double the price, double the preformance? Maybe, maybe not, maybe better compilers are the answer, maybe not, maybe we should just get another SGI. It's a total mess of a decision here. And, right now, since the real "cheifs" (pointy hair bosses) are making the decisions of who to even gets to talk, it get's even tougher.
I tell you, Dell not supporting more graphic cards, and not supporting Dual will play heavy against them in our decision. At the same time, "loyalty" plays it's ugly hand in favor of Dell for no good reason.
Some times, even when it looks like good news at first glance, it only makes the situation worse. The "Pointy Hair Boss" syndrome seems to play more heavy into these decisions than anything Dell or VAResearch can do/say/support. It's all well and good as far as Linux PR goes. But when you need the box NOW, and avaliability of Dell's limited support is coming "in a few weeks," and you can't even get the pointy haired bosses to TALK to the suppliers, this news just makes me wanna scream.
Sorry, I think I am babbleing incoherently now, I'll shut up, I just had to vent. Larry, if your out there, I am still fighting for you guys here, I hope I get them to pick up the damn phone. (Damn the phone, which BTW, Dell isn't answering today when they called to get details about the Dual's and Linux. Strike TWO for Dell.)
You may want to Check thier Pricing/Configuration before you say it's out of line. I see a heck of a lot of hardware there for under 1/2 of the $50,000 you mention.
For example, in December VA began shipping the VArServer 4100 system, a Red Hat Linux server that features four 450-MHz Intel Pentium II Xeon processors with 1GB of RAM on each, upgradable to 2GB. Pricing starts at $49,995.
1GB of RAM _ON_ each? What? I do believe they meant to say 1M full speed Cache Ram on each. If they stuck 1G on a CPU, they would be like world record holders or something....
I disagree, see my previous post one level up, same level as yours. I know my way around a box, and I can spark a shell and insert a kernel module on my own during install. But when they are rejected, the whole thing is fuxered.... Not just some little package that needs to be updated.
So, I just grabed an old SuSE 5.1 boot disk, and it worked flawlessly (and on ONE disk, yes, a ONE FLOPPY, NO CDROM complete install). Of course, then I had to upgrade a ton of packages.
I have a great deal of respect for Red Hat, and I wish them and Bob Young the best of luck. But, putting out 3-5 versions a year should NOT be a priority, getting it right should. 4.2 was good. Ever since that, I feel they have released to frequently, and too soon. IMHO 6.0 should have Gnome 1.0, and Kernel 2.2.x, and should be blanket tested in a widely distributed prerelease version before shipping. But, I guess I can't complain too much, 80% of my boxes run Red Hat :-/
I don't fear Red Hat becoming an "Evil Empire at all. It's almost impossable. I would LOVE to see IBM sell a distribution (especially on thier G3 boxen!). I don't think that any of them can be a treat when the source is GPL'ed. The problem comes for the lack of a standard, and that is currently being addressed by the LSB committee (still working slowly and quietly in the background).
The only real threat would be that with the lack of something like the LSB, if some distribution grows too big, they BECOME the standard, (for example, you _know_ people would port to IBM's or Microsoft's version of Linux), and what if the big name distribution is non standard, leaving apps only working on that distribution.
Long live the LSB.
dearth of applications that can exploit
new SMP capabilities.
It's under the Linux Negitive side at the bottom of the page. Is there some exploit that screws up SMP Linux boxes running 2.2.0 or something? Or, is this his flowery way of saying that there aren't many things that really use SMP yet?
Believing it's the latter, isn't that totall BS because
Yes, not my subject, I could be totally wrong, but, I didn't think that that statement was totally true.
Nevermind, I got it, all, saved it local, and now can read it, and show to others... I would mirror, but I don't think it's legql.
I think they took it down, I get "no longer exists" now. Anyone have it in thier cache, I would like to read the whole thing.
You do realize that doing ALL that html on a subject of cool cases, when wooden cases do exist, is idiotic, when the only thing that even HINTS at a joke is the URL? Man, if you think it's funny, I have a joke for you... "Jefferson and Hamlton walked into a pub, Jefferson said free the slaves, and no one said a word.... They were all in the army..." (if you don't get it, read some history).
I could be wrong, but I believe that it would be very difficult to get a SMP board that does over 4 CPU's, but using PC104's or SlotPC's, you can probably fit around 20 full systems in one case if you really tried.
100BaseT would be the bottleneck, your right, and you would want at least 100baseT and a good hub IN the case also. It's tight, but it's not a bad idea. I would LOVE to actually see one (love to build one even more, but I lack the financial resources to do such a thing).
Which SGI case can you think of that is huge enought, and avaliable cheap? I am thinking Personal Iris, they should be damn cheap, but do you think it's big enough?
Take the biggest Full Tower case you can find, and gut out all the bracketing, and make custom brackets to put in as many PC104's as you can fit (like this one). Then, use something like a 16M solid state IDE drive for each of the PC104's to put like
You might be able to squeze in like 10/20 full computers into one case with this idea, and use 100baseT to connect them, NFS to a few big SCSI drives in the case, and have it completely self contained! You probably would be limited to 300MHz socket 7 CPU's, but, w/ 10-20 of them in one case, working together the Beowulf way, it might totally rock.
Problems? Sure, money, custom case, and cooling!
This is the dumbest fake I have ever seen? When the html is soo anoying you cna't even read the text because of the layers, how are most people suppose to figure out it's a fake?
Wooden cases ARE made, by a few people, including IE NET, and I find them a fairly cool thing, in thier own way. So, I don't see the humor in spoofing something that is real. I mean, it's not a stupid idea, it's a cool idea. And it's out there.
Someone might have to explain what is suppose to be funny about this to me... Like, would it be funny if I put up a site that said I was building and selling MIPS servers, and did elaborate html, with photos and details, but I didn't really make or sell them? I don't think it's funny, I just don't get it. It's stupid.
index.html was ok, I find it anoying to redirect to home.html, because it makes going "back" a tad more difficult, but then some crap blocked me from reading the home.html, so I clicked some button I didn't even know what was, and ended up in "products." Fine, it's a mess, but it looked like the product might be interesting. I started reading it, and all of a sudden I couldn't see anything, and my load was spiked, graphics were dumped all over the place in the browser window, slowly moving, and MY POINTER started "jumping" when I tried to get to the "back" button! I can run scripts in gimp, two or three at a time on several images, and have NEVER seen that happen!
This site should DEFINATELY be linked on a "what not to do" page about html design.
So, what do you know about IA-32/Linux/SGI? What do you suspect the "server" design will be based on?
"As we enter the low-end of the market with 32-bit systems, we will support Windows NT and Linux, again giving our customers what they have been asking for," said Vrolyk.
I read comp.sys.sgi.hardware and the other comp.sys.sgi.* groups often... Do you? What do YOU see people asking for? SGI/LINUX servers? or SGI/Linux w/ X + GL?
PS. I think there may be an update by Monday, but I am not sure.
I have also read many other SGI comments and stories lately, and the burning question is X. Everyone is asking about X. So, lot's of people care about X. And when SGI announces that they will offically support linux for servers, people want to know "what about X?"
What's so hard to understand about people wanting X? Linux already runs on the Indy and the VPC, and therefore there is nothing new about being able to have a Linux/SGI server. The news story is about "SGI Offically Supporting Linux" and thier statement is that they will support it on SERVERS... Well, great, afaik, the support we recieve for IRIX is NOT great, and VERY expensive, so, although it's good PR for LINUX and SGI, it's not really that exciting of news when it comes down to the facts. What IS interesting about it is, with "offical" SGI support for linux in "some" ways, is this going to lead to what people REALLY want from SGI/LINUX... which is X.
What's so hard to understand about asking the _real_ question burning in peoples minds?
This is the goal of the CAJUN (Car Audio Jukebox UNIX) project, isn't it? :-)
So, how far away are the SGI propriatry X drivers for the VPC now? Will GL be included? Assuming that SGI writes closed source drivers for X in Linux, is this a good or a bad thing? (good to have useable new hardware with superior graphics to existing x86 Linux, bad cause it might be closed?)
http://gnustep.current. nu/comments.php3?sid=99/01/27/125608 is the correct link for the story.
IMHO Educational institutions should do it the way UofM's Engineering Department did. With an equal budjet, you could server far more users and/or provide better computing power to all users. (considering probably 9 out of 10 workstations usually carry a light load at most times anyway..... but even still....)
They aren't my posts, but I thought they were interesting, so I decided to share them with you ;-)
If it weren't for that, I would be happy if Red Hat did NOT come out with a new distribution just because of 2.2.0, and just put it in the updates. Then, they could push/wait for Gnome 1.0, and get rid of the awful default FVWM95 crud, and release "6.0" which would be cool. But... It's no doubt a quite complex issue, do you wait a little while and "get it all right" or do you jump on the bandwagon and try to be "The First" to do a 2.2.x distribution?