Clearly "POSO" is not a real company that could place a $500,000 order. Note also that the email address on the site is a hotmail address, and hotmail can't be used for business purposes. Sounds very much like a puppet company set up by LinuxOne folks to create fake demand. Shouldn't they be in their quiet period by now anyways? --JRZ
I actually like Dr. Dobbs' Journal quite a bit. Sure, it's full of ads, but I like their ads. Occasionally spotty quality (and one or two VB-related pieces!) don't overwhelm the good stuff that's still there. --JRZ
I use it on and off when I want that "professional" CDE feel. It is very much like the CDE, but also pretty quick and light. You also get the benefit of GTK themes, if you're into that sort of thing. The default configuration and icons, however, reveal how old-school the primary developer is. It's interesting to see someone who still considers xclock and xcalc to be among their most important applications. . . The default icons are mostly very simple 8-bit.xpms, but you can replace them trivially (the panel is very easy to configure). Its xftree file manager doesn't come up automatically in the default config, but it's really pretty good as well, considering that it's not meant to be the same sort of thing as a gmc or kfm. If you think that KDE and GNOME's taskbars are too "windows-like", the XFCE panel might be more your thing. I use 3.1.2 and the stability and speed are very good. --JRZ
Inprise was woth less than $300 million this morning. It really would've been nice to see free versions of VisiBroker, Delphi, C++ builder, Interbase, etc. . . Not that I think this would happen in a million years. BTW, in terms of stock, Cygnus was also valued at 6 times as much as Motorola paid for Metrowerks. While I understand the reasons for Red Hat's purchase, doesn't it seem like they overpaid a bit ? --JRZ
Java v1.3, which is actually a solid client-side programming possibility. v1.2 is nice for servers and networking, but Be is really looking for client applications. --JRZ
Well, now Cobalt (having shipped 17,000 servers) is kicking the crap out of SGI in terms of market cap. Those guys really should put out a high-end Linux server appliance anways. Cobalt/Rebel.com have not even come close to sewing up the market, and the SGI boys have more than enough engineering expertise. --JRZ
Last time this subject came up, a number of people lamented the idea that KDE would abandon the "open standard" of CORBA for a new approach. I'm glad Mosfet has cleared things up a bit by explaining that libICE is equally a standard which, in fact, is a lot more common in the Unix/Linux world than CORBA. What I find really strange is the argument that GNOME's use of CORBA makes it more standards compliant than KDE. Don't get me wrong, if ORBit works well for this application, that's fantastic, keep using it. But ORBit actually implements only a very small fraction of a modern CORBA standard (MICO is fully 2.2 compliant with all the bells and whistles, so it is a slug in terms of performance), and does not yet provide C++ or Java bindings. Essentially, it's a handy, GNOME-only solution, like KParts is a handy, KDE-only solution. --JRZ
that large websites would all post the equivalent of a colophon. You know, an explanation of how and with what the site is built: backend programming methods (EJB, NSAPI, mod_perl), design tools, web server, database, etc. It's always fascinating to see a little bit of a publisher's design philosophies in the colophon for a book, and I think that goes double for a web site. --JRZ
I can certainly see them making it "free" like IE to capture the market, with fairly generous source code licensing to manufacturing partners (it's not that hard to get the WinCE compared to, say, the Win2000 source code). This could help convince some manufacturers, but I think that opening the source would be a little dangerous. It would be too easy to clone the APIs and the interesting parts of the system. Of course, it would force Palm to, basically, do the same thing, just as they hit Netscape. They'd also run into possible "dumping" laws in the antitrust world, since it would be SUCH a transparent attempt to cut off the PalmOS's lifeline. --JRZ
This is absolutely ridiculous. Mr. Perens is claiming that Kaffe is "300% faster" than some competitors without giving concrete benchmarks or even stating who those competitors are. Obviously you can create a loaded scenario by pitting Kaffe with JIT vs. Blackdown with JIT disabled. Even worse is the AC who claimed that Kaffe could be 2x as fsat as HotSpot. For real, 3rd party benchmarks, check out the Volano report at: http://www.javaworld.com/javaworld/jw-10-1999/jw -10-volano.html In terms of speed, Kaffe rated a 389, while IBM's JDK got a 1770. Yup, the IBM kit was more than 4 times as fast. On top of that, Transvirtual doesn't pass Java compatibility tests, so if you develop on their VM, you have no reason to believe that your code will work on a real JVM. Open source is not necessarily equal to best of breed. --JRZ
Check out this Java World article for real benchmarks for a variety of JDKs on Intel hardware. The scalability (page 2) is both interesting and disappointing. The numbers on the Blackdown and Kaffe JVMs really show how badly a new JVM is needed for Linux. --JRZ
>Just as any person should be able to make a >decision/carry out the wishes of an individual's >euthenasia. eg. cases of terminal illness. These >are huge ethical decisions but should be owned by >the people whose lives they most directly effect. But clearly it MOST directly affects the life of the newborn, who has no say at all in the decision! If parents have some right to decide that it's right for their infant to be euthanized, and we grant them that right based on the fact that the child can not yet make its own decisions, what, then do we do with a four-year-old? Clearly he/she is too young to decide what to do with his/her life. So can a parent euthanize a bratty toddler to "save the world some suffering"? --JRZ
A number of comments have either questioned whether Linux is superior to the existing alternatives or pointed out that the superior tech does not always win in marketing battles. Both statements have validity, but neither really gets to the core of the dilemma. Linux is not disruptive because it's a radical, new, great technology (it's not). It's disruptive because (a) It works well for reasonable-sized market (small servers) and (b) It's unbelieveably cheap and easy to advance (due to the GPL) compared to existing technologies. Put 'em together, and you have necessary and sufficient conditions for a disruptive technology, as they say. --JRZ
If a server is mysteriously shut down (due to crash, hardware, or power failure), an OS with a good JFS won't lose any data or have to fsck like Linux does after a bad shutdown. If every byte is mission critical, or if you have a large database that would take hours to fsck, this is a must-have. --JRZ
Try "Corporate Level Strategy." I can't remember the author's name. It's about multi-business corporations and GE is a major case study. They do a lot of compartmentalization of their different businesses, allowing them to grow at their own pace. It's not a bad book about corporations that have grown a lot by acquisition, but it's not targetted at the brightest bulbs in the box, if you know what I mean. Aimed at big-company CEOs. --JRZ
A great deal of academic/government research ends up as open source. Take Beowulf, for example, or some interesting ideas like the Reiserfs filesystem. Part of the problem here is that the academic CS community and the OSS community aren't one in the same (any more). So good ideas from one side don't always flow to the other. I know of several really great research projects in everything from compilers to OS that have been released under the GPL from my school, but I don't exactly hear about them on Freshmeat. --JRZ
"Where is the OS/390 desktop environment?" "Where is the SCO video editing software?" "Is there a decent Mathematica-clone for Netware?" "Call Fujitsu and get theirs, or wait for Borland's (already ported) or Metrowerks' (coming very soon) compilers." None of these is a good reason not to use these operating systems in a server environment. I would also never argue that commercial software, that Windows software is inherently bad, or that open source is a magical solution to all our problems. Instead, I argued that Linux is a key, low-end server platform that will be (and already has been) disruptive in the operating system industry. I agree that most Linux distributors have very, very poor QA for the software they distribute. That's part of the opportunity that exists for established vendors (hardware, OS, applications, whatever) with very high QA standards to enter the market before Red Hat or Caldera really takes a solid hold. --JRZ
Someone doesn't seriously want to go to school to learn CFML and ASP, do they?!??!?!?? If they do, I have a perfect school for them: it's called a Wednesday afternoon with two books "Teach yourself ASP in 21 days" and "Teach yourself CFML in 21 days." Excellent university. Total cost: 10 hours and $100. Otherwise, it's sad but true. Most credible schools want you to learn the foundations of computer science as well as the applications. --JRZ
This should NOT be listed as a critque of the Cathedral and the Bazaar. It's a much deeper analysis of the open source development model. I did, however, really enjoy some of his shots at C&B. I know that that paper is important for historical reasons (e.g. its importance in the Mozilla decision), but otherwise it's never struck me as a particularly interesting piece. Arguments by analogy are inherently weak; even if the argument is airtight, one can always attack the link between the real world situation and the metaphor. That's partly what this author did by demonstrating that the "bazaar" image is not always appropriate for OSS projects. --JRZ
Man, Idealab patents everything! Business models, algorithms, the wheel, whatever. It's basically the cornerstone of their whole corporate philosophy. On the other hand, they could put the patent under the following license: "Patent X44236845kj is freely licensed for use in any piece of software which is covered by the GNU Public License. Any entity wishing to license the patent for use in a non-GPL piece of software should contact the patent holder for pricing." Then the GPL'ed software serves as a great proof-of-concept, as well as a roadmap of how to implement the concept in a real system. Not that I expect them to do this, but it would be an interesting decision. --JRZ
A company called ExOffice (www.exoffice.com) is pretty far along in this. Their product will be called Intalio. Contains all you mentioned plus a little bit of an XML development environment. --JRZ
Why are internet transactions special that they can't be taxed? A sale is a sale is a sale, and adding a 1% sales tax would NOT suddenly devastate the internet economy. --JRZ
If you're a big PHP fan, you should check out Midgard (at www.midgard-project.org). It provides some sweet content management and administration features (very similar to some of Zope's capabilities), but is based on a slightly-modified version of PHP. They have experimental ODBC support, and that should be solidified very soon. --JRZ
Old, resonably uninteresting news
on
SGI Releases IDE
·
· Score: 1
Hey, it's great that SGI is releasing this sort of thing, but they did it weeks ago, and Jessie has already been announced on half a dozen different sites. So far, it's just a front-end to gdb and gprof, and many of those (DDD, KDbg, etc) exist already. If they're right about its being a real framework for a serious IDE (class browser, syntax highlighting, makefile management), then I'll be interested. Anybody working in C/C++ should check out KDevelop, the best IDE for Unix I've seen yet. --JRZ
Well done piece, but has anyone been thinking otherwise? Their failure to support Java or server-side apps on Linux has been apparent for a long time. If platforms really did become irrelevant due to the wide availability of great JVMs and Java software, wouldn't that push most low-to-mid-range users towards free, stable platforms on commodity hardware? Not exactly what Sun wants. Of course, IBM will put out their Java2 JVM for Linux soon enough, and then we can really kick things up. --JRZ
Slashdot Mirror
Clearly "POSO" is not a real company that could place a $500,000 order. Note also that the email address on the site is a hotmail address, and hotmail can't be used for business purposes. Sounds very much like a puppet company set up by LinuxOne folks to create fake demand. Shouldn't they be in their quiet period by now anyways? --JRZ
I actually like Dr. Dobbs' Journal quite a bit. Sure, it's full of ads, but I like their ads. Occasionally spotty quality (and one or two VB-related pieces!) don't overwhelm the good stuff that's still there. --JRZ
I use it on and off when I want that "professional" CDE feel. It is very much like the CDE, but also pretty quick and light. You also get the benefit of GTK themes, if you're into that sort of thing. The default configuration and icons, however, reveal how old-school the primary developer is. It's interesting to see someone who still considers xclock and xcalc to be among their most important applications. . . The default icons are mostly very simple 8-bit .xpms, but you can replace them trivially (the panel is very easy to configure). Its xftree file manager doesn't come up automatically in the default config, but it's really pretty good as well, considering that it's not meant to be the same sort of thing as a gmc or kfm. If you think that KDE and GNOME's taskbars are too "windows-like", the XFCE panel might be more your thing. I use 3.1.2 and the stability and speed are very good. --JRZ
Inprise was woth less than $300 million this morning. It really would've been nice to see free versions of VisiBroker, Delphi, C++ builder, Interbase, etc. . . Not that I think this would happen in a million years. BTW, in terms of stock, Cygnus was also valued at 6 times as much as Motorola paid for Metrowerks. While I understand the reasons for Red Hat's purchase, doesn't it seem like they overpaid a bit ? --JRZ
Java v1.3, which is actually a solid client-side programming possibility. v1.2 is nice for servers and networking, but Be is really looking for client applications. --JRZ
Well, now Cobalt (having shipped 17,000 servers) is kicking the crap out of SGI in terms of market cap. Those guys really should put out a high-end Linux server appliance anways. Cobalt/Rebel.com have not even come close to sewing up the market, and the SGI boys have more than enough engineering expertise. --JRZ
Last time this subject came up, a number of people lamented the idea that KDE would abandon the "open standard" of CORBA for a new approach. I'm glad Mosfet has cleared things up a bit by explaining that libICE is equally a standard which, in fact, is a lot more common in the Unix/Linux world than CORBA.
What I find really strange is the argument that GNOME's use of CORBA makes it more standards compliant than KDE. Don't get me wrong, if ORBit works well for this application, that's fantastic, keep using it. But ORBit actually implements only a very small fraction of a modern CORBA standard (MICO is fully 2.2 compliant with all the bells and whistles, so it is a slug in terms of performance), and does not yet provide C++ or Java bindings. Essentially, it's a handy, GNOME-only solution, like KParts is a handy, KDE-only solution.
--JRZ
that large websites would all post the equivalent of a colophon. You know, an explanation of how and with what the site is built: backend programming methods (EJB, NSAPI, mod_perl), design tools, web server, database, etc.
It's always fascinating to see a little bit of a publisher's design philosophies in the colophon for a book, and I think that goes double for a web site.
--JRZ
I can certainly see them making it "free" like IE to capture the market, with fairly generous source code licensing to manufacturing partners (it's not that hard to get the WinCE compared to, say, the Win2000 source code). This could help convince some manufacturers, but I think that opening the source would be a little dangerous. It would be too easy to clone the APIs and the interesting parts of the system. Of course, it would force Palm to, basically, do the same thing, just as they hit Netscape.
They'd also run into possible "dumping" laws in the antitrust world, since it would be SUCH a transparent attempt to cut off the PalmOS's lifeline.
--JRZ
This is absolutely ridiculous. Mr. Perens is claiming that Kaffe is "300% faster" than some competitors without giving concrete benchmarks or even stating who those competitors are. Obviously you can create a loaded scenario by pitting Kaffe with JIT vs. Blackdown with JIT disabled. Even worse is the AC who claimed that Kaffe could be 2x as fsat as HotSpot.w -10-volano.html
For real, 3rd party benchmarks, check out the Volano report at:
http://www.javaworld.com/javaworld/jw-10-1999/j
In terms of speed, Kaffe rated a 389, while IBM's JDK got a 1770. Yup, the IBM kit was more than 4 times as fast. On top of that, Transvirtual doesn't pass Java compatibility tests, so if you develop on their VM, you have no reason to believe that your code will work on a real JVM.
Open source is not necessarily equal to best of breed.
--JRZ
Check out this Java World article for real benchmarks for a variety of JDKs on Intel hardware. The scalability (page 2) is both interesting and disappointing. The numbers on the Blackdown and Kaffe JVMs really show how badly a new JVM is needed for Linux. --JRZ
>Just as any person should be able to make a >decision/carry out the wishes of an individual's >euthenasia. eg. cases of terminal illness. These >are huge ethical decisions but should be owned by >the people whose lives they most directly effect.
But clearly it MOST directly affects the life of the newborn, who has no say at all in the decision! If parents have some right to decide that it's right for their infant to be euthanized, and we grant them that right based on the fact that the child can not yet make its own decisions, what, then do we do with a four-year-old? Clearly he/she is too young to decide what to do with his/her life. So can a parent euthanize a bratty toddler to "save the world some suffering"?
--JRZ
A number of comments have either questioned whether Linux is superior to the existing alternatives or pointed out that the superior tech does not always win in marketing battles. Both statements have validity, but neither really gets to the core of the dilemma.
Linux is not disruptive because it's a radical, new, great technology (it's not). It's disruptive because (a) It works well for reasonable-sized market (small servers) and (b) It's unbelieveably cheap and easy to advance (due to the GPL) compared to existing technologies.
Put 'em together, and you have necessary and sufficient conditions for a disruptive technology, as they say.
--JRZ
If a server is mysteriously shut down (due to crash, hardware, or power failure), an OS with a good JFS won't lose any data or have to fsck like Linux does after a bad shutdown. If every byte is mission critical, or if you have a large database that would take hours to fsck, this is a must-have.
--JRZ
Try "Corporate Level Strategy." I can't remember the author's name. It's about multi-business corporations and GE is a major case study.
They do a lot of compartmentalization of their different businesses, allowing them to grow at their own pace.
It's not a bad book about corporations that have grown a lot by acquisition, but it's not targetted at the brightest bulbs in the box, if you know what I mean. Aimed at big-company CEOs.
--JRZ
A great deal of academic/government research ends up as open source. Take Beowulf, for example, or some interesting ideas like the Reiserfs filesystem. Part of the problem here is that the academic CS community and the OSS community aren't one in the same (any more). So good ideas from one side don't always flow to the other. I know of several really great research projects in everything from compilers to OS that have been released under the GPL from my school, but I don't exactly hear about them on Freshmeat.
--JRZ
"Where is the OS/390 desktop environment?"
"Where is the SCO video editing software?"
"Is there a decent Mathematica-clone for Netware?"
"Call Fujitsu and get theirs, or wait for Borland's (already ported) or Metrowerks' (coming very soon) compilers."
None of these is a good reason not to use these operating systems in a server environment.
I would also never argue that commercial software, that Windows software is inherently bad, or that open source is a magical solution to all our problems.
Instead, I argued that Linux is a key, low-end server platform that will be (and already has been) disruptive in the operating system industry.
I agree that most Linux distributors have very, very poor QA for the software they distribute. That's part of the opportunity that exists for established vendors (hardware, OS, applications, whatever) with very high QA standards to enter the market before Red Hat or Caldera really takes a solid hold.
--JRZ
Someone doesn't seriously want to go to school to learn CFML and ASP, do they?!??!?!?? If they do, I have a perfect school for them: it's called a Wednesday afternoon with two books "Teach yourself ASP in 21 days" and "Teach yourself CFML in 21 days." Excellent university. Total cost: 10 hours and $100.
Otherwise, it's sad but true. Most credible schools want you to learn the foundations of computer science as well as the applications.
--JRZ
This should NOT be listed as a critque of the Cathedral and the Bazaar. It's a much deeper analysis of the open source development model. I did, however, really enjoy some of his shots at C&B. I know that that paper is important for historical reasons (e.g. its importance in the Mozilla decision), but otherwise it's never struck me as a particularly interesting piece. Arguments by analogy are inherently weak; even if the argument is airtight, one can always attack the link between the real world situation and the metaphor. That's partly what this author did by demonstrating that the "bazaar" image is not always appropriate for OSS projects. --JRZ
Man, Idealab patents everything! Business models, algorithms, the wheel, whatever. It's basically the cornerstone of their whole corporate philosophy. On the other hand, they could put the patent under the following license:
"Patent X44236845kj is freely licensed for use in any piece of software which is covered by the GNU Public License. Any entity wishing to license the patent for use in a non-GPL piece of software should contact the patent holder for pricing."
Then the GPL'ed software serves as a great proof-of-concept, as well as a roadmap of how to implement the concept in a real system.
Not that I expect them to do this, but it would be an interesting decision.
--JRZ
A company called ExOffice (www.exoffice.com) is pretty far along in this. Their product will be called Intalio. Contains all you mentioned plus a little bit of an XML development environment. --JRZ
Why are internet transactions special that they can't be taxed? A sale is a sale is a sale, and adding a 1% sales tax would NOT suddenly devastate the internet economy.
--JRZ
If you're a big PHP fan, you should check out Midgard (at www.midgard-project.org). It provides some sweet content management and administration features (very similar to some of Zope's capabilities), but is based on a slightly-modified version of PHP. They have experimental ODBC support, and that should be solidified very soon.
--JRZ
Hey, it's great that SGI is releasing this sort of thing, but they did it weeks ago, and Jessie has already been announced on half a dozen different sites. So far, it's just a front-end to gdb and gprof, and many of those (DDD, KDbg, etc) exist already. If they're right about its being a real framework for a serious IDE (class browser, syntax highlighting, makefile management), then I'll be interested. Anybody working in C/C++ should check out KDevelop, the best IDE for Unix I've seen yet. --JRZ
Well done piece, but has anyone been thinking otherwise? Their failure to support Java or server-side apps on Linux has been apparent for a long time. If platforms really did become irrelevant due to the wide availability of great JVMs and Java software, wouldn't that push most low-to-mid-range users towards free, stable platforms on commodity hardware? Not exactly what Sun wants.
Of course, IBM will put out their Java2 JVM for Linux soon enough, and then we can really kick things up.
--JRZ