Future generations...
on
Is Linux Dead?
·
· Score: 4, Interesting
...will point to this time and say "2002 was the year Microsoft lost the war."
Why will they say this?
A free office suite has become available.
Microsoft is raising prices for Office and the OS while the market slumps.
Mozilla turns 1.0.
Lindows goes mainstream.
The sentencing phase of the trial is complete.
The avalanche of private civil suits begins.
But then again, I still don't understand why SQL Server is selling so well when the same codebase can be obtained for free from linux.sybase.com.
Still, free software is a flood that is rising around Microsoft, and Microsoft is busy trying to build something that floats. It is unlikely that they will succeed, given the importance of their legacy support.
If you really believe that the key to success is an uninformed and uncaring fanbase, be my guest, give it a try.
I actually bought 18 last night. After listening to 3 tracks, I am unimpressed.
Moby's major influences (as he confesses) are Christianity and Veganism. Now we all know that it is easy to find revolutionary and profound ideas in these subjects even today.
However, from what I've heard of 18, it is neither revolutionary nor profound.
18 is Moby's attempt to rest on his laurels from Play.
Rather that slash and burn Alpha, HP would have made all of its customers far more happy by agreeing to take EV8 to silicon. I'm sure this was in the realm of the possible.
Yes, cross-license with Intel up the wazoo and sell your employees to Intel if you like, but deliver to your customers what they need to keep their datacenters for the next decade, and also bring a stunning and seminal SMT product to market.
While we're on the subject, unifying HP-UX and Tru64 into a "TruHP" might have scored a few notches on the cluestick. Let's face it: a lot of things about HP-UX just plain suck (especially the packaging system, as Tru64 announced it was moving to RPM). HP is just beginning to implement dynamic kernel tunables and even their whole enterprise file system is outsourced. I am totally underwhelmed. When they lose the performance edge, I will have no sentimental attachment to this kludge.
Just like IBM and Sequent, HP has knifed products that work for products that don't. May Opteron be the undoing of you all.
All Linux distributions need Kazaa-like installers
on
Slackware 8.1 is Released
·
· Score: 3, Insightful
Download a bootnet floppy or static Linux executible which checks a list of mirrors, tests bandwidth to find the fastest, and downloads the ISOs and/or does your install.
RedHat up2date seems to use such a mechanism; download times off this network are much faster than updates.redhat.com.
I screwed up my main Linux system this weekend, and hunting for a fast mirror on win98 is annoying.
I am not a physicist, but I seem to remember a discussion of FTL properties implied by work by Einstein-Podolsky-Rosen and extended by J.S. Bell.
The particular discussion I remember involved the emission of paired electrons that travel some arbitrary distance. When the path of one of these electrons is altered by some external force, the path of the paired electron is also instantaneously diverted (with some limitation due to frame of reference) in an equal and opposite direction, implying an FTL transfer of information, if not of matter or energy. As I remember, this has an intense impact on the assertion of local causality. Google seems to turn up a large number of references on this subject.
In any case I do not understand why practical applications of this phenomena have not yet been developed.
Please pardon any inadequacies of this summary; I am a lowly engineer and computer scientist entirely unqualified to be commenting on such lofty subjects.
On odd occasions, I do need a large ISO image right away, but usually if I'm downloading an ISO for Solaris, Oracle, or Red Hat, I can wait up to several days before I really need it.
Cable companies should set up special proxies for downloading large files, and distribute multi-platform download applications that use them. These proxies should automatically throttle bandwidth consumption in periods of high-utilization.
Cable companies with bandwidth problems should provide credits for people who use such mechanisms, and surcharges for people who use straight ftp.
It would also be nice if Kazaa et al provided standard http/ftp interfaces for download, for the pirates among us.
As a long time fan of Red Hat, I have a few questions:
Will UnitedLinux include proprietary code, or will the entire source be available? What happens to YAST? Will the installer source be open?
If most or all of the source is open and the product is popular, then you must assume that independent binary distributions of UnitedLinux will appear. Do you have any plans to legally restrict these activities?
There also seems to be some confusion (from Suse) as to what is open and what is not. Are all the vendors on the same page? Is there any infighting?
Do you plan to license any interesting proprietary technology for UnitedLinux? For example, might UnitedLinux license Apple's current incarnation of NextStep, or are you planning a Cocoa-compatible version of GNUStep for UnitedLinux (excuse this line of thought; I just bought a Mac)? Are you planning any proprietary additions to UnitedLinux which might give you an edge over your (more open) competetors (who are currently beating the tar out of you in market share)?
Along these same lines, have you had any conversations with Apple concerning the cross-licensing of your respective technologies? Access to their GUI components would be a boon to you, and Apple is making its first attempts at a server appliance and might find some UnixWare components to be of great value (clustering, true System-V source, etc.). You might add Cocoa in the same way that Motif/CDE RPMS used to be added for a surcharge, and even this might drive sales.
I'm not familiar with the process, but will UnitedLinux pursue POSIX certification?
Yeah, you could make some money with this, but it will require a bit of fancy footwork.
As a cluster guru, I am curious about your take on database server clustering in both the commercial and the open-source space.
First, it appears that IBM DB2 has been wiping the floor with Oracle on the TPC benchmarks lately, and Oracle "RAC" has been a flop. However, IBM is not using any hardware from its proprietary server lines, but instead relies on clusters of "federated" databases running on 32 standard PCs running either Linux or Windows. It does appear that Oracle still generally beats IBM in raw performance on a single system (as IBM refuses to post any non-clustered benchmarks AFAIK).
Do you think that any of the hype over either of these vendors cluster packages is worth attention? Do you agree with Sun's claim that TPC(-C) no longer has any practical relevance? It all seems to be getting rather silly.
Second, is there any push to make any of the ACID-leaning open databases (Postgres, SAP-DB, etc.) fault-tolerant, perhaps using Mosix? I assume this would require modifications to Postgres enabling it to access raw partitions. Have you had any talks with the Red Hat Database people about cluster modifications to Postgres, just out of curiousity?
AMD's Opteron (Hammer) will integrate Non-Uniform Memory Access (NUMA) features. IBM has had significant NUMA design experience through its own products and acquisitions.
Does IBM plan any major NUMA efforts for Linux? Is there any synergy with NUMA-Q? How will any NUMA efforts impact IBM's Itanium commitments? Any possibility that we might see an Opteron port of AIX 5L?
And, perhaps most importantly, if AMD's NUMA efforts prove fruitful, might IBM be forced to de-emphasize it to protect its competing product lines? After all, AIX only recently became "partitionable", while Solaris has had this feature for some time...
Stanford (or any real univeristy) doesn't have a "Java" course. It has courses on data structures which might use Java.
You mean like course number cs193J "Programming in Java" being taught this summer? (http://cs.stanford.edu/Courses/Schedules/2001-200 2.summer.html) Perhaps Stanford is no longer a "real university" - shall we observe a moment of silence to mourn its passing?
Any schools that don't have a big football team? I bet DeVry is on there too!
If membership in the Oracle Academic Initiative belies a desperate, cash strapped department, then will you please explain why every big-ten school is listed under the OAI directory?
This is all a question of independent review of academic standards. Java instruction from Stanford is not the same as Java instruction from MIT is not the same as Java instruction from UIUC et al. With the academic adoption of JCert, these programs would be the same. This is a valuable goal.
Currently, colleges and universities are free to adopt their own CS standards as long as they are not ridiculously lax. Enforced certifications in Java and SQL would tighten things up considerably, and they should be implemented today.
What is mainly wrong with CS in higher education is the lack of standards. It is hard to implement standards, and your typical academic attitude belies the laziness of the tenure system in general.
What magical quality is there with rhetoric 101 in a 300-seat lecture hall that will teach a person to think? How can you argue that such an experience is worthwhile?
The most valuable things I learned in my (comparatively challenging) academic career were discovered outside of the classroom. Then, it was the only game in town. Now, it is not.
College is no longer blindly accepted as a touchstone of personal scholarship. As an academic, prove your worth, or lose what little respect that you still have.
I'm going to talk about DB certifications, because that interests me the most.
Right now, basic IBM DB2 certification is free of charge. The program isn't very well-organized (compared to Oracle OCP), but you can't argue with the cost. It will also get you some basic SQL skills.
The free qualifier is only available in June, August, October, and November of this year. Information on it is available at http://certify.torolab.ibm.com.
Perhaps more important than this, however, is that both IBM and Oracle have programs for integrating their certs into college courses (Oracle exams are even half-price for students, IIR). The URLs for their academic sites are:
What I recommend to you is that you demand that your institution of higher learning participate in vendor partnership programs such as these. Such partnerships a)indicate that the institution is interested in imparting practical and industry-relevant experience to you, and b)ensure you of a higher starting salary than those unfortunate individuals attending more institutions with less focus on your needs (and more on the ease of their tenured professors).
I might especially recommend JCert. If your college teaches Java, they ought not to be afraid to have graduates from their programs independently certified. Any hesitation on this point belies a lack of faith in the quality of their own instruction.
This sort of thing is new ground for most schools, and I think if your school is willing to at least let you work these certs as independent study, then they should still be considered. However, I've seen a few cases now where administration stonewalls (University of Iowa), even though they are a member of the program. You might think about making your entire tech-elective track nothing but certs.
So go for the campus tour, nod and smile at their spiel, then negotiate hard and in writing that these certs will be accepted as tech electives.
And don't be afraid to remind them that the ROI for certs is far greater than for college tuition in the short and even medium term.
If Sun went to the maintainer of GNU tar and said "integrate these patches and we will use your app as the primary Solaris TAR," how quickly do you think the GNU people would wet themselves? They'd leap at the chance.
What exactly is it about SysV cp, mv, tar, awk, ls et al that makes them so much more valuable than their GNU equivalents?
Sun has no idea how to address Linux. However, if Sun were to replace all possible SysV components in Solaris with their GNU equivalents, they would be much farther down the road towards a free OS than the Sun Community Source License ever got them. This would at least give them some short-term PR, plus cutting development costs.
I really don't understand why every UNIX distribution isn't making these moves. If I were to say that 90% of the GNU UNIX utilities could replace the proprietary components with no visible effect to the OS, would that be a conservtative estimate?
Sun could go even further by wrapping Red Hat Linux around the Solaris kernel, and scaling Red Hat onto an e15k.
And, if Sun were to take the step of open-sourcing the Solaris kernel, Sun could put an end to the question of enterprise UNIX on any Intel platform - Sun takes all.
Come on, guys, wake up! You're asleep at the wheel!
HP-UX is profoundly weak in several areas. If Carly had come out and said "the new TruHP UNIX will unify the strengths of HP-UX and Digital OSF/1 UNIX on ia64," I would have thought her to be much more reasonable.
Really, the tru64 kernel should simply replace the HP-UX kernel, with the important addition of Veritas support.
Now would also be a good time to redesign the software packaging mechanisms and implement something like RedHat up2date.
But instead, HP throws us the same old trash. I hope their market share continues to erode.
As an example, the condemnation by the Commercial Court of Nanterre, France, on 27th September 2001 of Microsoft Corp. to a penalty of 3 million francs in damages and interest, for violation of intellectual property (piracy, to use the unfortunate term that your firm commonly uses in its publicity).
Assuming that a sufficient quantity of the material could be synthesized with the required tensile strength to hold a human, it might not have that much mass.
Slashdot Mirror
...will point to this time and say "2002 was the year Microsoft lost the war."
Why will they say this?
But then again, I still don't understand why SQL Server is selling so well when the same codebase can be obtained for free from linux.sybase.com.
Still, free software is a flood that is rising around Microsoft, and Microsoft is busy trying to build something that floats. It is unlikely that they will succeed, given the importance of their legacy support.
If you really believe that the key to success is an uninformed and uncaring fanbase, be my guest, give it a try.
I actually bought 18 last night. After listening to 3 tracks, I am unimpressed.
Moby's major influences (as he confesses) are Christianity and Veganism. Now we all know that it is easy to find revolutionary and profound ideas in these subjects even today.
However, from what I've heard of 18, it is neither revolutionary nor profound.
18 is Moby's attempt to rest on his laurels from Play.
Moby, get off your a**.
According to Daniel K. Benjamin's "Oracle 9i New Features For Administrators Exam Guide," Oralce 9i introduces:
Oracle has a lot of problems, but standards conformance is not one of them. Oracle is one of the few databases to have certified with NIST for SQL-92.
As is documented at the end of today's article here.
Rather that slash and burn Alpha, HP would have made all of its customers far more happy by agreeing to take EV8 to silicon. I'm sure this was in the realm of the possible.
Yes, cross-license with Intel up the wazoo and sell your employees to Intel if you like, but deliver to your customers what they need to keep their datacenters for the next decade, and also bring a stunning and seminal SMT product to market.
While we're on the subject, unifying HP-UX and Tru64 into a "TruHP" might have scored a few notches on the cluestick. Let's face it: a lot of things about HP-UX just plain suck (especially the packaging system, as Tru64 announced it was moving to RPM). HP is just beginning to implement dynamic kernel tunables and even their whole enterprise file system is outsourced. I am totally underwhelmed. When they lose the performance edge, I will have no sentimental attachment to this kludge.
Just like IBM and Sequent, HP has knifed products that work for products that don't. May Opteron be the undoing of you all.
Download a bootnet floppy or static Linux executible which checks a list of mirrors, tests bandwidth to find the fastest, and downloads the ISOs and/or does your install.
RedHat up2date seems to use such a mechanism; download times off this network are much faster than updates.redhat.com.
I screwed up my main Linux system this weekend, and hunting for a fast mirror on win98 is annoying.
I am not a physicist, but I seem to remember a discussion of FTL properties implied by work by Einstein-Podolsky-Rosen and extended by J.S. Bell.
The particular discussion I remember involved the emission of paired electrons that travel some arbitrary distance. When the path of one of these electrons is altered by some external force, the path of the paired electron is also instantaneously diverted (with some limitation due to frame of reference) in an equal and opposite direction, implying an FTL transfer of information, if not of matter or energy. As I remember, this has an intense impact on the assertion of local causality. Google seems to turn up a large number of references on this subject.
In any case I do not understand why practical applications of this phenomena have not yet been developed.
Please pardon any inadequacies of this summary; I am a lowly engineer and computer scientist entirely unqualified to be commenting on such lofty subjects.
On odd occasions, I do need a large ISO image right away, but usually if I'm downloading an ISO for Solaris, Oracle, or Red Hat, I can wait up to several days before I really need it.
Cable companies should set up special proxies for downloading large files, and distribute multi-platform download applications that use them. These proxies should automatically throttle bandwidth consumption in periods of high-utilization.
Cable companies with bandwidth problems should provide credits for people who use such mechanisms, and surcharges for people who use straight ftp.
It would also be nice if Kazaa et al provided standard http/ftp interfaces for download, for the pirates among us.
As a long time fan of Red Hat, I have a few questions:
Yeah, you could make some money with this, but it will require a bit of fancy footwork.
As a cluster guru, I am curious about your take on database server clustering in both the commercial and the open-source space.
First, it appears that IBM DB2 has been wiping the floor with Oracle on the TPC benchmarks lately, and Oracle "RAC" has been a flop. However, IBM is not using any hardware from its proprietary server lines, but instead relies on clusters of "federated" databases running on 32 standard PCs running either Linux or Windows. It does appear that Oracle still generally beats IBM in raw performance on a single system (as IBM refuses to post any non-clustered benchmarks AFAIK).
Do you think that any of the hype over either of these vendors cluster packages is worth attention? Do you agree with Sun's claim that TPC(-C) no longer has any practical relevance? It all seems to be getting rather silly.
Second, is there any push to make any of the ACID-leaning open databases (Postgres, SAP-DB, etc.) fault-tolerant, perhaps using Mosix? I assume this would require modifications to Postgres enabling it to access raw partitions. Have you had any talks with the Red Hat Database people about cluster modifications to Postgres, just out of curiousity?
AMD's Opteron (Hammer) will integrate Non-Uniform Memory Access (NUMA) features. IBM has had significant NUMA design experience through its own products and acquisitions.
Does IBM plan any major NUMA efforts for Linux? Is there any synergy with NUMA-Q? How will any NUMA efforts impact IBM's Itanium commitments? Any possibility that we might see an Opteron port of AIX 5L?
And, perhaps most importantly, if AMD's NUMA efforts prove fruitful, might IBM be forced to de-emphasize it to protect its competing product lines? After all, AIX only recently became "partitionable", while Solaris has had this feature for some time...
There is enough information in my previous post to warrant a plug here.
Basically, IBM DB2 certification is free, and Oracle tests and materials are available for half price if you are in an academic program.
This was linked off Red Hat's site sometime back in the RHCE hype. This data is pretty old; I haven't seen them post any updates.
I myself have an OCP; I would think Java would be more reasonable cert for a new grad.
http://www.crn.com/sections/special/ssurvey/ssurvDidn't I read recently that SCO released the rights to the V7 source? Was it just the kernel or the whole OS?
How difficult would it be to target this for x86? How difficult would it be to make it run XFree? Would the kernel need to be updated for POSIX?
You mean like course number cs193J "Programming in Java" being taught this summer? (http://cs.stanford.edu/Courses/Schedules/2001-200 2.summer.html) Perhaps Stanford is no longer a "real university" - shall we observe a moment of silence to mourn its passing?
A master of evasion you are not.
Suicide's not worth it, dude.
Also, let's have some objective statistics on the subject:
http://www.crn.com/sections/special/ssurvey/ssur ve y01.asp?ArticleID=25726
If membership in the Oracle Academic Initiative belies a desperate, cash strapped department, then will you please explain why every big-ten school is listed under the OAI directory?
This is all a question of independent review of academic standards. Java instruction from Stanford is not the same as Java instruction from MIT is not the same as Java instruction from UIUC et al. With the academic adoption of JCert, these programs would be the same. This is a valuable goal.
Currently, colleges and universities are free to adopt their own CS standards as long as they are not ridiculously lax. Enforced certifications in Java and SQL would tighten things up considerably, and they should be implemented today.
What is mainly wrong with CS in higher education is the lack of standards. It is hard to implement standards, and your typical academic attitude belies the laziness of the tenure system in general.
What magical quality is there with rhetoric 101 in a 300-seat lecture hall that will teach a person to think? How can you argue that such an experience is worthwhile?
The most valuable things I learned in my (comparatively challenging) academic career were discovered outside of the classroom. Then, it was the only game in town. Now, it is not.
College is no longer blindly accepted as a touchstone of personal scholarship. As an academic, prove your worth, or lose what little respect that you still have.
I'm going to talk about DB certifications, because that interests me the most.
Right now, basic IBM DB2 certification is free of charge. The program isn't very well-organized (compared to Oracle OCP), but you can't argue with the cost. It will also get you some basic SQL skills.
The free qualifier is only available in June, August, October, and November of this year. Information on it is available at http://certify.torolab.ibm.com.
Perhaps more important than this, however, is that both IBM and Oracle have programs for integrating their certs into college courses (Oracle exams are even half-price for students, IIR). The URLs for their academic sites are:
http://oai.oracle.com/
http:///www-3.ibm.com/software/data/highered/
Other good certification websites:
http://suned.sun.com
http://www.jcert.org
http://www.cisco.com
What I recommend to you is that you demand that your institution of higher learning participate in vendor partnership programs such as these. Such partnerships a)indicate that the institution is interested in imparting practical and industry-relevant experience to you, and b)ensure you of a higher starting salary than those unfortunate individuals attending more institutions with less focus on your needs (and more on the ease of their tenured professors).
I might especially recommend JCert. If your college teaches Java, they ought not to be afraid to have graduates from their programs independently certified. Any hesitation on this point belies a lack of faith in the quality of their own instruction.
This sort of thing is new ground for most schools, and I think if your school is willing to at least let you work these certs as independent study, then they should still be considered. However, I've seen a few cases now where administration stonewalls (University of Iowa), even though they are a member of the program. You might think about making your entire tech-elective track nothing but certs.
So go for the campus tour, nod and smile at their spiel, then negotiate hard and in writing that these certs will be accepted as tech electives.
And don't be afraid to remind them that the ROI for certs is far greater than for college tuition in the short and even medium term.
If Sun went to the maintainer of GNU tar and said "integrate these patches and we will use your app as the primary Solaris TAR," how quickly do you think the GNU people would wet themselves? They'd leap at the chance.
What exactly is it about SysV cp, mv, tar, awk, ls et al that makes them so much more valuable than their GNU equivalents?
Sun has no idea how to address Linux. However, if Sun were to replace all possible SysV components in Solaris with their GNU equivalents, they would be much farther down the road towards a free OS than the Sun Community Source License ever got them. This would at least give them some short-term PR, plus cutting development costs.
I really don't understand why every UNIX distribution isn't making these moves. If I were to say that 90% of the GNU UNIX utilities could replace the proprietary components with no visible effect to the OS, would that be a conservtative estimate?
Sun could go even further by wrapping Red Hat Linux around the Solaris kernel, and scaling Red Hat onto an e15k.
And, if Sun were to take the step of open-sourcing the Solaris kernel, Sun could put an end to the question of enterprise UNIX on any Intel platform - Sun takes all.
Come on, guys, wake up! You're asleep at the wheel!
HP-UX is profoundly weak in several areas. If Carly had come out and said "the new TruHP UNIX will unify the strengths of HP-UX and Digital OSF/1 UNIX on ia64," I would have thought her to be much more reasonable.
Really, the tru64 kernel should simply replace the HP-UX kernel, with the important addition of Veritas support.
Now would also be a good time to redesign the software packaging mechanisms and implement something like RedHat up2date.
But instead, HP throws us the same old trash. I hope their market share continues to erode.
What was this about?
Assuming that a sufficient quantity of the material could be synthesized with the required tensile strength to hold a human, it might not have that much mass.
But granted, it does strain credulity.