I have been working in higher ed (not K-12) application software for > 10 years. Every consortium for administrative software that I have experienced - and there have been several - have been colossal failures. Consortiums are the ultimate design by committee, and DBC is the worst way to design software.
I'd love to see some open source limited market application software projects. There aren't many, for this reason: This type of application software needs domain experts and programmers. Ideally, if the domain expert is also a programmer, you'll get the best software. A talented programmer who is also a domain expert and wants to build open source software is a rare bird indeed. Add to that the immense commitment required for all but the simplest application programs, and you see why limited market application programs are not open source.
According to Ronald Kessler, author of The Bureau: The Secret History of the FBI, Freeh is also responsible for the failure of the FBI to keep up with technology. At the end of Freeh's tenure, FBI agents were using 486-class computers and had to email attachments to home computers to transmit pictures. Freeh resisted upgrading the FBI mainframe infrastructure as well. He's clearly not capable of making judgments about crypto.
I don't know what NASA uses to compile those images, but I do know that the GIMP can read and write FITS, the standard format for astronomical images. Many of the Hubble images, in FITS format, are available from the links listed in the other posts.
For amateur astrophotography, some of which rivals the NASA shots, much of the image processing is done with Photoshop. Most of the tools used to process those images are also available in the GIMP.
Here's a quote from a Sun marketing person in charge of J2EE branding:
...having a strong brand and compatibility standards are important to the development of a robust market for J2EE platform products, tools, and components. The J2EE Compatible brand has achieved significant momentum over the past two years, and we want to make sure that any open source efforts don't impact the viability of that effort.
My translation: The J2EE brand is like the little Windows flag on the Windows box -- you gotta pay the man to get it.
I have 7.2 with mozilla and galeon. I installed the mozilla 0.9.6 (--nodeps) and galeon segfaults. So it looks like the latest galeon binary is not compatible with 0.9.6 which is pretty much the way it always works with galeon and mozilla.
I don't know of a production quality open source OLAP database. That said, my company is working on a basic set of OLAP class libraries and tools (in perl and php) which enable summarization and reporting of data warehouses. It is not an end user tool, but we've been using it to build query applications on dimensional data marts. Our plan all along has been to release it as open source.
We have not released yet for two reasons. First, it is not ready yet -- we've seen a lot of alpha quality open source projects languish on Freshmeat because they're just not usable. We want to include enough sample code and documentation to make it a decent starting point for a programmer to build an OLAP web application. Second, this post is the first real interest we've seen in open source OLAP, so we weren't in any real hurry.
BTW, we use MySQL, because it is fast and reliable for read-only data.
This is probably related to this description of the NuSphere product on the MySQL site:
NuSphere MySQL Advantage includes an unauthorized modified version of MySQL which includes support for GEMINI tables. As long as GEMINI is not released under open source, as required by the GPL license of the MySQL server, we at MySQL AB can't recommend anyone to use this distribution. Also, NuSphere uses our trademark in the product name and elsewhere without our permission.
How is it that KDE can put out software that is arguably as good as Gnome without millions of dollars of venture capital?
It seems that the past year's (impressive) Gnome improvements have been contributed in large part by Ximian and Eazel, whereas KDE has been able to create some nice software without an equivalent source of funding. Is there a lesson somewhere in the KDE project for Ximian and Eazel?
My (small) company creates web-based applications and delivers them as an ASP. We use Linux, Apache, MySQL and Perl or PHP (LAMP), and we're happy with the results. Our software costs approach zero, and we don't have to worry about per-seat or per-server licensing.
On occasion, a feature we need is missing from the LAMP toolset. This feature is available in a Microsoft product, and we think about switching. However, when we look at the economics of the situation, we find a way to work around our "need" for the feature.
I'm no Linux zealot -- I've been in this business for 15 years, and I've used (and been happy with) Microsoft products. But, as a small software shop, the business decision to use LAMP instead of Windows 2000, IIS, SQL Server and ASP seems like a no-brainer.
My question: How does Microsoft compete with LAMP in this kind of environment?
I'm a web developer (php, perl, mysql) also, and I use a Linux box running a Windows emulator. I've used VMWare and Win4Lin as the emulator. Here's the breakdown:
Win4Lin Pluses: Cheaper and Faster than VMWare.
Win4Lin Minuses: Must patch the kernel (or install patched kernel from RPM)
VMWare Pluses: Can run OSes other than Win9X in VMWare Pro (e.g., Linux on Linux to test installations). No kernel change required.
VMWare Minuses: Slower than Win4Lin and more expensive.
I use Win4Lin now, mainly to check pages in IE. My IDE of choice is Emacs.
I occasionally click ads on Slashdot or News.com; hardly ever on Yahoo, Salon, etc. Why? Because Slashdot is targeted to some of my interests.
Given the opportunity to specify interests for all OSDN banner ads, I'd do it. I think it would increase click-thru in addition to letting you shop around a targeted demographic. Seems like a win-win. Bring it on.
We are read-only ~10 million row database in MySQL on a dual PIII 800 w/1GB RAM. Indexed lookups are fast (sub second). MySQL was created as a data warehousing database, and when you use it for its original design purpose, it is fast.
We found that the limiting factor is the number of indexes on each table. All indexes for each table are stored in one file. As you add indexes, reindex performance degrades; we're guessing that the cause is b-tree rebalancing. Anyway, we addressed that issue by creating and indexing sub tables. It's a query-only application, so we can afford to do things like that.
Also, if you read through the thread on news.kde.org you'll see that TheKompany will not be open sourcing the groupware server. This is is the way they intend to make money.
My personal experience with wrist pain (and if you read the sore hands site, you'll find that there are may kinds of wrist/joint pain that are not carpal tunnel) is that it can be fixed by:
Good ergonomics, especially keyboard placement.
Better awareness of other stresses in your life. Check out The Mind Body Prescription. It is a book by a respected, practicing physician who points out how your repressed anger and stress can contribute to (back/wrist/etc.) pain. Literally, psychosomatic illness. I thought it was bullshit until I was desperate enough to try it, and it made things a lot better.
Vitamins: nothing radical. A good multivitamin, a Vitamin B supplement, and SAMe. This is the #1 prescribed anti-arthritic in Germany, and is sold in the US as a nutritional supplement in places like GNC. It also supposedly has anti-depressive properties. A lot of studies show that increasing one of the B vitamins helps in Carpal Tunnel.
The bottom line is that, about 1.5 years ago, I couldn't work at a keyboard for more than 15 minutes without a lot of pain, numbness, etc. Now, I don't think about it -- working 12 hour days is no issue. I am definitely not a person who immediately goes to alternative therapies. Actually, I was seeing a MD and physical therapist for over a year before trying these three things out of total frustration.
BTW, I fully admit that #3 may be a placebo. #2 may be also. So what. They worked for me.
I've built medium-size web applications (6 month development projects) in both, and settled on a mix of PHP for presentation and Perl for scripting. They are both great languages. Here's what I've found from experience:
PHP was built specifically as a web presentation language. It has everything you need, and no more, to build the pieces of a web site that a user touches.
PHP also runs leaner in Apache. With mod_perl, you are embedding a perl interpreter and all of the memory used by your apps into each Apache instance. There is some shared memory going on, but mod_perl is much bigger. The programmer must also be much more careful about memory usage with mod_perl. Perl advocates would argue that mod_perl is therefore faster, and they might be right. But PHP seems fast enough.
Perl is, IMHO, a better general-purpose language. PHP does have a command-line version, but Perl just "feels better" to me to write scripts to load databases, munge files, etc.
A huge part of the Perl benefit is the extensibility provided by CPAN modules. CPAN is a beautiful thing. PHP does not have anything like it that I've found. Most PHP code sharing is.inc modules that you need to customize to get to work. There's a lot of great code out there for PHP, don't get me wrong, it's just harder to get at and use than CPAN.
Finally, both Perl and PHP are rock-solid. When there's an error in a web app I write, it's mine, and that's a huge blessing.
Neither MySQL nor Interbase 6.0 were tested in TPC-C. MySQL doesn't have enough SQL92 conformance. Interbase 6.0's ODBC driver isn't ready yet. They tried to use Interbase 5.0 but couldn't get it to work.
The test where all databases competed is the AS3AP test. A little more research shows that this is a test with mixed updates and retrievals. MySQL 3.22 is known to have poor performance with a large number of mixed updates and retrievals. This may explain why the MySQL line peaked and then fell off for this test.
If you ever have to do high-speed packet capture, the packet filter (BPF) in the BSD stack is far superior to the Linux one (even the latest kernels, where packet capture has been improved).
Wasn't SETI@Home getting more data than they could use because a lot more people were sending data than they initially anticipated? Won't this card just make that situation (much) worse?
The other factor is people. It's easy to say "well, you're this big company with all this money, throw more developers at it" Even forgetting the fact that "more doesn't always equal better, or faster", I don't care if you are AOL, Microsoft, IBM, or whoever...finding developers skilled enough to work with a task that complex is next to impossible in this industry.
This is an important point. I'll bet that a lot of the problem with Mozilla is brain drain. JWZ is one major example. Mozilla is turning in to a "death march" project, and in this employment environment, nobody has to work on this kind of project.
Slashdot Mirror
I have been working in higher ed (not K-12) application software for > 10 years. Every consortium for administrative software that I have experienced - and there have been several - have been colossal failures. Consortiums are the ultimate design by committee, and DBC is the worst way to design software.
I'd love to see some open source limited market application software projects. There aren't many, for this reason: This type of application software needs domain experts and programmers. Ideally, if the domain expert is also a programmer, you'll get the best software. A talented programmer who is also a domain expert and wants to build open source software is a rare bird indeed. Add to that the immense commitment required for all but the simplest application programs, and you see why limited market application programs are not open source.
According to Ronald Kessler, author of The Bureau: The Secret History of the FBI, Freeh is also responsible for the failure of the FBI to keep up with technology. At the end of Freeh's tenure, FBI agents were using 486-class computers and had to email attachments to home computers to transmit pictures. Freeh resisted upgrading the FBI mainframe infrastructure as well. He's clearly not capable of making judgments about crypto.
I don't know what NASA uses to compile those images, but I do know that the GIMP can read and write FITS, the standard format for astronomical images. Many of the Hubble images, in FITS format, are available from the links listed in the other posts.
For amateur astrophotography, some of which rivals the NASA shots, much of the image processing is done with Photoshop. Most of the tools used to process those images are also available in the GIMP.
I have 7.2 with mozilla and galeon. I installed the mozilla 0.9.6 (--nodeps) and galeon segfaults. So it looks like the latest galeon binary is not compatible with 0.9.6 which is pretty much the way it always works with galeon and mozilla.
Opera and Mozilla are blocked at noon EST.
I don't know of a production quality open source OLAP database. That said, my company is working on a basic set of OLAP class libraries and tools (in perl and php) which enable summarization and reporting of data warehouses. It is not an end user tool, but we've been using it to build query applications on dimensional data marts. Our plan all along has been to release it as open source.
We have not released yet for two reasons. First, it is not ready yet -- we've seen a lot of alpha quality open source projects languish on Freshmeat because they're just not usable. We want to include enough sample code and documentation to make it a decent starting point for a programmer to build an OLAP web application. Second, this post is the first real interest we've seen in open source OLAP, so we weren't in any real hurry.
BTW, we use MySQL, because it is fast and reliable for read-only data.
Anyone know when/where RPMS will be released?
This is probably related to this description of the NuSphere product on the MySQL site:
NuSphere MySQL Advantage includes an unauthorized modified version of MySQL which includes support for GEMINI tables. As long as GEMINI is not released under open source, as required by the GPL license of the MySQL server, we at MySQL AB can't recommend anyone to use this distribution. Also, NuSphere uses our trademark in the product name and elsewhere without our permission.
The "elsewhere" presumably includes mysql.org.
This is a serious question, not a troll:
How is it that KDE can put out software that is arguably as good as Gnome without millions of dollars of venture capital?
It seems that the past year's (impressive) Gnome improvements have been contributed in large part by Ximian and Eazel, whereas KDE has been able to create some nice software without an equivalent source of funding. Is there a lesson somewhere in the KDE project for Ximian and Eazel?
My (small) company creates web-based applications and delivers them as an ASP. We use Linux, Apache, MySQL and Perl or PHP (LAMP), and we're happy with the results. Our software costs approach zero, and we don't have to worry about per-seat or per-server licensing.
On occasion, a feature we need is missing from the LAMP toolset. This feature is available in a Microsoft product, and we think about switching. However, when we look at the economics of the situation, we find a way to work around our "need" for the feature.
I'm no Linux zealot -- I've been in this business for 15 years, and I've used (and been happy with) Microsoft products. But, as a small software shop, the business decision to use LAMP instead of Windows 2000, IIS, SQL Server and ASP seems like a no-brainer.
My question: How does Microsoft compete with LAMP in this kind of environment?
I'm a web developer (php, perl, mysql) also, and I use a Linux box running a Windows emulator. I've used VMWare and Win4Lin as the emulator. Here's the breakdown:
Win4Lin Pluses: Cheaper and Faster than VMWare.
Win4Lin Minuses: Must patch the kernel (or install patched kernel from RPM)
VMWare Pluses: Can run OSes other than Win9X in VMWare Pro (e.g., Linux on Linux to test installations). No kernel change required.
VMWare Minuses: Slower than Win4Lin and more expensive.
I use Win4Lin now, mainly to check pages in IE. My IDE of choice is Emacs.
I occasionally click ads on Slashdot or News.com; hardly ever on Yahoo, Salon, etc. Why? Because Slashdot is targeted to some of my interests.
Given the opportunity to specify interests for all OSDN banner ads, I'd do it. I think it would increase click-thru in addition to letting you shop around a targeted demographic. Seems like a win-win. Bring it on.
I'm running the latest Flash plugin in 0.8 and it works fine. It worked fine in 0.7 also, btw.
We are read-only ~10 million row database in MySQL on a dual PIII 800 w/1GB RAM. Indexed lookups are fast (sub second). MySQL was created as a data warehousing database, and when you use it for its original design purpose, it is fast.
We found that the limiting factor is the number of indexes on each table. All indexes for each table are stored in one file. As you add indexes, reindex performance degrades; we're guessing that the cause is b-tree rebalancing. Anyway, we addressed that issue by creating and indexing sub tables. It's a query-only application, so we can afford to do things like that.
while ($story_needed) {
@idiots = find_idiots($senseless_yakkers);
foreach $moron (@idiots) {
push(@comments,get_quotes($moron));
}
$story = intersperse_speculation(@comments);
print html_format($story);
}
Also, if you read through the thread on news.kde.org you'll see that TheKompany will not be open sourcing the groupware server. This is is the way they intend to make money.
Check out the roadmap. Groupware is coming, with shared calendar targeted for March.
- Good ergonomics, especially keyboard placement.
- Better awareness of other stresses in your life. Check out The Mind Body Prescription . It is a book by a respected, practicing physician who points out how your repressed anger and stress can contribute to (back/wrist/etc.) pain. Literally, psychosomatic illness. I thought it was bullshit until I was desperate enough to try it, and it made things a lot better.
- Vitamins: nothing radical. A good multivitamin, a Vitamin B supplement, and SAMe. This is the #1 prescribed anti-arthritic in Germany, and is sold in the US as a nutritional supplement in places like GNC. It also supposedly has anti-depressive properties. A lot of studies show that increasing one of the B vitamins helps in Carpal Tunnel.
The bottom line is that, about 1.5 years ago, I couldn't work at a keyboard for more than 15 minutes without a lot of pain, numbness, etc. Now, I don't think about it -- working 12 hour days is no issue. I am definitely not a person who immediately goes to alternative therapies. Actually, I was seeing a MD and physical therapist for over a year before trying these three things out of total frustration.BTW, I fully admit that #3 may be a placebo. #2 may be also. So what. They worked for me.
I've built medium-size web applications (6 month development projects) in both, and settled on a mix of PHP for presentation and Perl for scripting. They are both great languages. Here's what I've found from experience:
.inc modules that you need to customize to get to work. There's a lot of great code out there for PHP, don't get me wrong, it's just harder to get at and use than CPAN.
PHP was built specifically as a web presentation language. It has everything you need, and no more, to build the pieces of a web site that a user touches.
PHP also runs leaner in Apache. With mod_perl, you are embedding a perl interpreter and all of the memory used by your apps into each Apache instance. There is some shared memory going on, but mod_perl is much bigger. The programmer must also be much more careful about memory usage with mod_perl. Perl advocates would argue that mod_perl is therefore faster, and they might be right. But PHP seems fast enough.
Perl is, IMHO, a better general-purpose language. PHP does have a command-line version, but Perl just "feels better" to me to write scripts to load databases, munge files, etc.
A huge part of the Perl benefit is the extensibility provided by CPAN modules. CPAN is a beautiful thing. PHP does not have anything like it that I've found. Most PHP code sharing is
Finally, both Perl and PHP are rock-solid. When there's an error in a web app I write, it's mine, and that's a huge blessing.
If you ever have to do high-speed packet capture, the packet filter (BPF) in the BSD stack is far superior to the Linux one (even the latest kernels, where packet capture has been improved).
Wasn't SETI@Home getting more data than they could use because a lot more people were sending data than they initially anticipated? Won't this card just make that situation (much) worse?
This is an important point. I'll bet that a lot of the problem with Mozilla is brain drain. JWZ is one major example. Mozilla is turning in to a "death march" project, and in this employment environment, nobody has to work on this kind of project.
I use a recent nightly build on RH 6.2. It does not "run great". It is slow and crashes a lot more than Netscape 4.7 on the same platform.