I have played around with this exact idea in the past. I got a semi-working prototype working, but then gave up as the huge amount of work needed became apparent. Here are the highlights:
Create a parallel configuration system depository somewhere outside of/etc, like/conf. That depository should only contain system and meta-system level stuff.
Keep application-specific configurations in the application directory.
Get rid of the ages old/usr,/bin, etc, etc. Blatantly copy OS X. OS X has solved most of the metadata management issues already.
Maintain a separate, read-only hierarchy where programs can still get their config files (/etc), libraries (/usr/lib, etc), etc.
Have a system deamon that reads the new metadata repositories and creates the legacy ones for the legacy parts of the sytem.
XML is not that bad if you are not creating it or reading by hand (but the cool part is that you can fix it by hand as a last resort). And it's scriptable as hell. Check out the Gnosis XML utilities in Python (disclaimer: I did contribute to this project).
XML_Objectify for example allows you to build a full object tree of the XML repository, complete with attached methods (on a per *tag* basis) and everything. Creating a/etc file from an objectified XML one is a cinch that way.
One common misconception: Python only forces you to use consistent identation *within the same code block* not the entire file or project. Which makes perfect sensem if you think about it.
Re:Perl, Python under .NET?
on
.NETly News
·
· Score: 3, Interesting
I would like to know the answer to that as well. I went looking for Visual Python earlier today and there's zero info (that I could find at any rate) on Active State's site on interoperability with the other Python implementations (cPython and Jython mostly). No word on the standard library (that has a few C extensions; how will those be managed in.NET?) or win32all and the Python-COM bindings.
As a python fan I had high hopes that Python would be the only language to bridge the JVM-CLR religious war and allow you to work in both.
It seems that ActiveState is just plugging in Python to VS, not compiling python to IL.
It makes sense for some consumers: I'd love to use single-play DVDs that I don't have to return: I am constantly on the road. Where would you rent a DVD from if you are a business traveller? open a Blockbuster account in every franchise in the country? return it where? What if you fly cross-country or overseas?
If this thing catches on, I'd be happy to buy 3-4 DVDs I'd like to watch at some point, carry it in my laptop bag, and catch them at a long flight or at any hotel that doesn't have cable or sat TV --like most European business hotels...
I find the claim dubious. Bigger than what kind of database? Wal-Mart if famous for tracking every single little thing about their supply chain. Most grocers or hypermart chains do the same. I can easily see, say A&P or Tesco or Carrefour having multi-TB DBs, even petabyte DBs.
Also, the size is not the only thing that defines a database installation: numbers of simultaneuous users or concurrent transactions, read or write access, ability to rollback, quality of service standards are way more important in my book (and also for most big companies). Part of the reason DBs in that size range are rare is exactly that current technology does not scale up to those levels while maintaining rollbacks, read-write and fast user response.
I like the Wayback machine, but to compare it to a proper database is ludicrous. EMC or Veritas will give you much more for their 100TBs of storage than 400 x85 PCs... instant backups for one and way larger MTBF.
100 TBs do not make the biggest DB ever. I am personally working on an 60-70TB ERP system that's also writeable; I am sure there are bigger systems out there (e.g. Wal-Mart's or GM's ERP systems come to mind).
A read-only DB containing highly-compressible text does not really make for a very challenging datamine. Just because it's on and about the Web and sexier than a stodgy ERP system should not make you overlook the real technology.
I am an aerospace engineer, and I fly on A300s once a week. I also used to be involved in air safety and preventive maintainance for a military aviation.
Notes:
1. Single engine failure during take-off is the single worst design condition for a twin-jet like an A-300.
2. Single engine failures during take-off are always taken into account for any passenger aircraft. A simple engine
failure cannot bring down a jetliner.
3. What can bring down a jetliner is the consequences of an engine failure: fire in the wing, explosion of the wing fuel tanks, compound failure of all redundant hydraulic systems, pylon failure (which would expose fuel lines), etc.
However, most of the above reasons are well-known. Take-off is the hardest flight region, and most eventualities are taken into account into designing these birds.
Further, a quick search of NTSB's online air crash info database, reveals no incident involving an A-300 and engine failure in the last 5 yrs. This is not typical if a design error is to be blamed.
Thus, it can be two things: either a failure of preventive maintainance or sabotage. The former is possible, due to the recent massive layoffs in the airline business, but unlikely: airlines usually don't fire skilled personnel, and when/if they do, maintainance personnel tend to over-perform during times of crises.
Please stop assuming that somehow corners are cut when designing airliners or that aero engineers sit around saying "lets use combustive materials for this one, shall we"? We know that we only get one chance to avoid fatalities. Airliners are routinely designed with huge safety margins, usually on top of the worst-ever-recorded conditions.
I mean, how can anybody argue with the notion that a Cathedral is somehow inferior to a Bazaar? We all know Bazaars where it's at, that's what people look at these days, and travel to Paris and Rome and places to see and marvel at. Hardly anybody stops by the Notre Damme.
It's also pretty clear that anarchy by design and design by anarchy work well. After all, open source has brought some exceptionally innovative technologies to IT consumers in the past few years. We now can finally parse flat text files with greater speeds and more flexibility than ever before! And we keep bug-compatibility to programs written for 1960s computers that can be outperformed by a wristwatch! Now, that's what I call technology! Object orientation? component programming? that's for wussies who can't code in C, sh, or perl!
Finally, how can traditional software businesses compete with the multi-level marketing scheme of proselytizing users that become testers and developers and finally evangelists? It's obvious that all great engineering and scientific endeavours have been benefitted by active recruitement and by popular opinion, not some arrogant dude's idea of what 'right' is.
After all, software is tantamount to *speech*, not machinery. It should be spoken and transmitted freely, not designed and crafted like some piece of steel.
Oh, yeah, there was something else, but I am sure the replies to this will fill you in... something about advocacy or something...
GSM in the states only sucks if you live in the States. I have a dual band GSM 900/1900 phone (the Nokia 8890) and live in Europe. I have excellent access whenever I visit the US (which is often) as I usually remain in metro areas.
Even things like retrieving voicemail and full SMS back home is fully supported.
As far as I am concerned, global roaming is here now. What is really missing though, is a) cheap roaming (the roaming charges are ridiculously expensive the world over), and b) cheap data access. My provider charges me for data access outside my 100 free minute plan, and data access while roaming, that's just out of the question, financially, not technically. 2.5G or 3G hopefully may solve that...
I have integrated VSS with Compuware Track Record. Compuware claim that is doable out of the box but after much research I eventually just went through a VB applet to do it.
After that painful experience, I fired up Python and talked to VSS through it. I still have a buncha classes that wrap around a VSS client pretty extensively, so that may lay the groundword for an interface to Bugzilla or something.
One other thing that I meant to look into but never got a chance: VSS provides a fairly extensive OLE interface which is callable through Python (and I am guessing Perl too). It should be simple enough to wrap around that to provide integration with Bugzilla or some such, or alternatively, to finally get a decent unix VSS client (the one that MS points to simply sucks).
Banner ads worsen the user experience; print ads do not. Regular banner ads take up my bandwidth and worsen my browser's responsivenss. So, yeah, I will use a proxy, no question.B ack when I had a 1.2M DSL line, I didn't care about ads that much. I only resorted to the Proxomitron when pop-ups and pop-unders became the vogue.
The industry needs to get smarter. Download ads after (not before) the content of the page has loaded. Stop annoying the user. A click-through ad is much better than a pop-under or a flashing fake Windows alert box any day of the week. Sponsor content instead of obstracting its delivery. Something.
This has to be a hoax. The picture on the site is terrible, and the details don't make much sense (will be able to play PC games? what PC games? how, if it's a locked down 2nd generation XBox?)
And "HomeStation"? I mean, com'on, you mean to tell me that MS will spend millions promoting the XBox brand and they will not use it for this (call it XBox TV or something?). I don't buy it.
I didn't say the article made sense, I was just answering the parent post's question...
And I don't see why my answer is nonsense (simplified maybe, but not wrong). If you worked for Gulfstream, maybe you can back it up, aero engineer to aero engineer.
* The "intensity" (read: energy) of a boom is proportional (roughly) to the speed of the aircraft and the angle of attack of the wing or fuselage.
* To lower the energy wasted in a sonic boom, you can either go slower (neah...) or lower the angle of attack. For a wing, this is kinda easy: either sweep it back (notice a how much further back a fighter's wings are than an airliner's?) or make it thinner (so that the cross-sectional angle of attack, so to speak, is less).
* For a fuselage it gets trickier: a fighter need only fit one person, and you can extend the nose long enough to lower the leading angle of attack. And you don't care about traling shocks or really shocks at all, because you're in a fighter. You're supposed to terrify people.
* But for a commercial jet, you will have to take care of both ends of the fuselage, and the only way is to make them longer, and have them taper out smoother. Look at the Concorde's absurdly long nose (so long, it has to be pivoted so that the pilots can see the runway at take-offs and landings) and its thin tail. Now, you know why they're there.
Supersonic business jets have always been possible. However, new, more efficient engines and cheaper high performance materials are only now making them affordable (well, relatively at least:-)...
You assume that Germany has a paging network. I do not know if it does, but it's not a safe assumption. Paging is NOT widespread in Europe. Cell phones have leapfrogged pagers here.
My solution and my suggestion would be a Palm with an IR-capable phone/w a builtin moded (dont forget the modem, IR is NOT enough). Most Nokia 6- and 8- series phones can fill that role. Then you can have IMAP and POP and web to your heart's content.
However, not all cell carriers charge for data calls the same (GSM can tell). My carrier doesnt count data calls as part of my 100 minutes i get free a month, even if i am calling a local DAN. Also, plan on about 9.6kbps on a good day. POP is fine, but web would be a bit of a stretch.
As for graffiti, I aint crazy about it either. Try a paste-on keyboard, see if that helps.
And also try T9 for your SMS messages (you need a T9 phone --most modern Nokias have it standard). For short messages w/ common words T9 is sometimes faster than qwerty...
No Cygwin; zsh for win32 deals with drive letters and path delimeters itself. The userland utilities are 90% native ports with only a couple that had to be linked against cygwin.dll. If you look around enough on the Net, you will find pretty much everything...
Admittedly, NT is not BSD, or OSX; you still have to think DOS (as for example, for drive letters), but overall it's very doable.
Don't get me wrong: I am excited about OSX, I used to be a NeXTStep zealot back in the day; however, win32 is viable as a development platform for Unix developers.
malamac@trammel:~> which python
c:/Python/python
malamac@trammel:~> find c:/Program\ Files/ -name Apache.exe -print
c:/Program Files/Apache Group/Apache/Apache.exe
malamac@trammel:~> net start mysql
The MySql service is starting.
The MySql service was started successfully.
Open Source needs detail-oriented, small-time, painstaking *work*. Not vision, not tirades against the Evil Empire of Redmond, not manifestos. It needs a lot of people to sit down and clean things up. Small things, things that collectively annoy the user, but individually are too boring for an OSS developer to bother with.
That's how good software, how good *anything* gets done. The OSS way however, has been that for something to be refined, polished, it has to "scratch an itch", it has to annoy someone enough *and* have that someone be skilled or talented enough to go fix it him/herself.
How do you get people to work on minute trivial things that they normally donot care about? the one time-tested way is, well, to *pay* them to do it. Where do you get the money? well, you will probably need a corporate structure that will fund these developers and that will be able to stand on its own two feet.
In other words you need companies, companies like Active State, theKompany, Digital Creations (now Zope) and a few (very few) others. You need to let companies sell OSS without bitching all the time that they are ruining your free lunch. You need to let companies have pay-to-play versions that are ahead of the OSS one so that they can financially support development, QA and documentation. You need to have companies have non-GPL licenses on their products without going Homeini on them.
Preferrably all of the above do not involve flammage and mail-bombing and invocations of Rights and Freedom. Writing software is not Speech, it is *not* equivalent to expressing one's opinion on common matters ("politics", people call them). Writing software is hard, painful and mostly boring work, engineering work. It needs good design, a painstaking devotion to quality and most of all someone to be paying that budget.
... I was modeling combustion during my undergrad years as far back as '93, and there were already several labs across the country doing computational combustion work. For a variety of reasons computational combustion is one of the hardest simulation fields you can tackle. Here's a plug to the leading lab of my old prof, the Computational Combustion Laboratory at Georgia Tech
Well, that's not quite true; you can argue that a language newbie will pick the most 'natural' way to write something in a given language. If that's true (and I know that's arguable in itself), then a newbie's benchmark is a pretty good comparison of how the languages stack up.
E.g.: I am a Python freak, but for some things, 'regular' Python idioms (like for loops) are slow, but functional Python features (like map()) are much faster. I would use map() but that would be cheating, 'coz I wouldn't know the same shortcut for Perl, for example...
A speculative answer since b-trees are my bread and butter (I am just now specing a 2TB data-mine): hundreds of thousands of entries (or hundreds of millions) should not really bother a b-tree. From the articles about Google, I am guessing they have implemented some sort of distributed b-tree app server, across all those COTS linux boxes.
I am curious as to what kind of implementation they are using; Google's roots would suggest some hacked form of Berkeley DB with lots of performance improvements.
Oh, well, just some guesswork... if I am close, I am expecting a job offer by the way:-)...
...Netscape will become Pathfinder for the GenX crowd. Only that when Pathfinder was live, only us "GenXers" (I hate that term too) were using the Web in the first place, and you know how Pathfinder turned out.
The browser war is dead anyway. The coming war will be Web-based services (think IM, stock quotes, money management); MS has the better technology, but AOL has the users... it will be an interesting battle:-)...
Slashdot Mirror
- Create a parallel configuration system depository somewhere outside of
/etc, like /conf. That depository should only contain system and meta-system level stuff.
- Keep application-specific configurations in the application directory.
- Get rid of the ages old
/usr, /bin, etc, etc. Blatantly copy OS X. OS X has solved most of the metadata management issues already.
- Maintain a separate, read-only hierarchy where programs can still get their config files (/etc), libraries (/usr/lib, etc), etc.
- Have a system deamon that reads the new metadata repositories and creates the legacy ones for the legacy parts of the sytem.
XML is not that bad if you are not creating it or reading by hand (but the cool part is that you can fix it by hand as a last resort). And it's scriptable as hell. Check out the Gnosis XML utilities in Python (disclaimer: I did contribute to this project). XML_Objectify for example allows you to build a full object tree of the XML repository, complete with attached methods (on a per *tag* basis) and everything. Creating aOne common misconception: Python only forces you to use consistent identation *within the same code block* not the entire file or project. Which makes perfect sensem if you think about it.
I would like to know the answer to that as well. I went looking for Visual Python earlier today and there's zero info (that I could find at any rate) on Active State's site on interoperability with the other Python implementations (cPython and Jython mostly). No word on the standard library (that has a few C extensions; how will those be managed in .NET?) or win32all and the Python-COM bindings.
As a python fan I had high hopes that Python would be the only language to bridge the JVM-CLR religious war and allow you to work in both.
It seems that ActiveState is just plugging in Python to VS, not compiling python to IL.
It makes sense for some consumers: I'd love to use single-play DVDs that I don't have to return: I am constantly on the road. Where would you rent a DVD from if you are a business traveller? open a Blockbuster account in every franchise in the country? return it where? What if you fly cross-country or overseas?
If this thing catches on, I'd be happy to buy 3-4 DVDs I'd like to watch at some point, carry it in my laptop bag, and catch them at a long flight or at any hotel that doesn't have cable or sat TV --like most European business hotels...
I hate to link a beta-level site from /., but that's exactly what I am trying out...
Hmmm... I can top that.
I find the claim dubious. Bigger than what kind of database? Wal-Mart if famous for tracking every single little thing about their supply chain. Most grocers or hypermart chains do the same. I can easily see, say A&P or Tesco or Carrefour having multi-TB DBs, even petabyte DBs.
Also, the size is not the only thing that defines a database installation: numbers of simultaneuous users or concurrent transactions, read or write access, ability to rollback, quality of service standards are way more important in my book (and also for most big companies). Part of the reason DBs in that size range are rare is exactly that current technology does not scale up to those levels while maintaining rollbacks, read-write and fast user response.
I like the Wayback machine, but to compare it to a proper database is ludicrous. EMC or Veritas will give you much more for their 100TBs of storage than 400 x85 PCs... instant backups for one and way larger MTBF.
100 TBs do not make the biggest DB ever. I am personally working on an 60-70TB ERP system that's also writeable; I am sure there are bigger systems out there (e.g. Wal-Mart's or GM's ERP systems come to mind).
A read-only DB containing highly-compressible text does not really make for a very challenging datamine. Just because it's on and about the Web and sexier than a stodgy ERP system should not make you overlook the real technology.
I am an aerospace engineer, and I fly on A300s once a week. I also used to be involved in air safety and preventive maintainance for a military aviation.
Notes:
1. Single engine failure during take-off is the single worst design condition for a twin-jet like an A-300.
2. Single engine failures during take-off are always taken into account for any passenger aircraft. A simple engine
failure cannot bring down a jetliner.
3. What can bring down a jetliner is the consequences of an engine failure: fire in the wing, explosion of the wing fuel tanks, compound failure of all redundant hydraulic systems, pylon failure (which would expose fuel lines), etc.
However, most of the above reasons are well-known. Take-off is the hardest flight region, and most eventualities are taken into account into designing these birds.
Further, a quick search of NTSB's online air crash info database, reveals no incident involving an A-300 and engine failure in the last 5 yrs. This is not typical if a design error is to be blamed.
Thus, it can be two things: either a failure of preventive maintainance or sabotage. The former is possible, due to the recent massive layoffs in the airline business, but unlikely: airlines usually don't fire skilled personnel, and when/if they do, maintainance personnel tend to over-perform during times of crises.
Please stop assuming that somehow corners are cut when designing airliners or that aero engineers sit around saying "lets use combustive materials for this one, shall we"? We know that we only get one chance to avoid fatalities. Airliners are routinely designed with huge safety margins, usually on top of the worst-ever-recorded conditions.
I mean, how can anybody argue with the notion that a Cathedral is somehow inferior to a Bazaar? We all know Bazaars where it's at, that's what people look at these days, and travel to Paris and Rome and places to see and marvel at. Hardly anybody stops by the Notre Damme.
It's also pretty clear that anarchy by design and design by anarchy work well. After all, open source has brought some exceptionally innovative technologies to IT consumers in the past few years. We now can finally parse flat text files with greater speeds and more flexibility than ever before! And we keep bug-compatibility to programs written for 1960s computers that can be outperformed by a wristwatch! Now, that's what I call technology! Object orientation? component programming? that's for wussies who can't code in C, sh, or perl!
Finally, how can traditional software businesses compete with the multi-level marketing scheme of proselytizing users that become testers and developers and finally evangelists? It's obvious that all great engineering and scientific endeavours have been benefitted by active recruitement and by popular opinion, not some arrogant dude's idea of what 'right' is.
After all, software is tantamount to *speech*, not machinery. It should be spoken and transmitted freely, not designed and crafted like some piece of steel.
Oh, yeah, there was something else, but I am sure the replies to this will fill you in... something about advocacy or something...
GSM in the states only sucks if you live in the States. I have a dual band GSM 900/1900 phone (the Nokia 8890) and live in Europe. I have excellent access whenever I visit the US (which is often) as I usually remain in metro areas.
Even things like retrieving voicemail and full SMS back home is fully supported.
As far as I am concerned, global roaming is here now. What is really missing though, is a) cheap roaming (the roaming charges are ridiculously expensive the world over), and b) cheap data access. My provider charges me for data access outside my 100 free minute plan, and data access while roaming, that's just out of the question, financially, not technically. 2.5G or 3G hopefully may solve that...
I have integrated VSS with Compuware Track Record. Compuware claim that is doable out of the box but after much research I eventually just went through a VB applet to do it.
After that painful experience, I fired up Python and talked to VSS through it. I still have a buncha classes that wrap around a VSS client pretty extensively, so that may lay the groundword for an interface to Bugzilla or something.
One other thing that I meant to look into but never got a chance: VSS provides a fairly extensive OLE interface which is callable through Python (and I am guessing Perl too). It should be simple enough to wrap around that to provide integration with Bugzilla or some such, or alternatively, to finally get a decent unix VSS client (the one that MS points to simply sucks).
Banner ads worsen the user experience; print ads do not. Regular banner ads take up my bandwidth and worsen my browser's responsivenss. So, yeah, I will use a proxy, no question.B ack when I had a 1.2M DSL line, I didn't care about ads that much. I only resorted to the Proxomitron when pop-ups and pop-unders became the vogue.
The industry needs to get smarter. Download ads after (not before) the content of the page has loaded. Stop annoying the user. A click-through ad is much better than a pop-under or a flashing fake Windows alert box any day of the week. Sponsor content instead of obstracting its delivery. Something.
I wanna test out this puppy, but there are only FTP mirrors listed and the firewall here isn't cooperating. Any HTTP mirrors?
This has to be a hoax. The picture on the site is terrible, and the details don't make much sense (will be able to play PC games? what PC games? how, if it's a locked down 2nd generation XBox?)
And "HomeStation"? I mean, com'on, you mean to tell me that MS will spend millions promoting the XBox brand and they will not use it for this (call it XBox TV or something?). I don't buy it.
I didn't say the article made sense, I was just answering the parent post's question...
And I don't see why my answer is nonsense (simplified maybe, but not wrong). If you worked for Gulfstream, maybe you can back it up, aero engineer to aero engineer.
Short lesson on high-speed aerodynamics follows:
:-)...
* The "intensity" (read: energy) of a boom is proportional (roughly) to the speed of the aircraft and the angle of attack of the wing or fuselage.
* To lower the energy wasted in a sonic boom, you can either go slower (neah...) or lower the angle of attack. For a wing, this is kinda easy: either sweep it back (notice a how much further back a fighter's wings are than an airliner's?) or make it thinner (so that the cross-sectional angle of attack, so to speak, is less).
* For a fuselage it gets trickier: a fighter need only fit one person, and you can extend the nose long enough to lower the leading angle of attack. And you don't care about traling shocks or really shocks at all, because you're in a fighter. You're supposed to terrify people.
* But for a commercial jet, you will have to take care of both ends of the fuselage, and the only way is to make them longer, and have them taper out smoother. Look at the Concorde's absurdly long nose (so long, it has to be pivoted so that the pilots can see the runway at take-offs and landings) and its thin tail. Now, you know why they're there.
Supersonic business jets have always been possible. However, new, more efficient engines and cheaper high performance materials are only now making them affordable (well, relatively at least
You assume that Germany has a paging network. I do not know if it does, but it's not a safe assumption. Paging is NOT widespread in Europe. Cell phones have leapfrogged pagers here.
/w a builtin moded (dont forget the modem, IR is NOT enough). Most Nokia 6- and 8- series phones can fill that role. Then you can have IMAP and POP and web to your heart's content.
My solution and my suggestion would be a Palm with an IR-capable phone
However, not all cell carriers charge for data calls the same (GSM can tell). My carrier doesnt count data calls as part of my 100 minutes i get free a month, even if i am calling a local DAN. Also, plan on about 9.6kbps on a good day. POP is fine, but web would be a bit of a stretch.
As for graffiti, I aint crazy about it either. Try a paste-on keyboard, see if that helps.
And also try T9 for your SMS messages (you need a T9 phone --most modern Nokias have it standard). For short messages w/ common words T9 is sometimes faster than qwerty...
No Cygwin; zsh for win32 deals with drive letters and path delimeters itself. The userland utilities are 90% native ports with only a couple that had to be linked against cygwin.dll. If you look around enough on the Net, you will find pretty much everything...
Admittedly, NT is not BSD, or OSX; you still have to think DOS (as for example, for drive letters), but overall it's very doable.
Don't get me wrong: I am excited about OSX, I used to be a NeXTStep zealot back in the day; however, win32 is viable as a development platform for Unix developers.
malamac@trammel:~> which python
c:/Python/python
malamac@trammel:~> find c:/Program\ Files/ -name Apache.exe -print
c:/Program Files/Apache Group/Apache/Apache.exe
malamac@trammel:~> net start mysql
The MySql service is starting.
The MySql service was started successfully.
malamac@trammel:~> echo $SHELL
d:\root\bin\zsh.exe
malamac@trammel:~> uname
Windows_NT
Open Source needs detail-oriented, small-time, painstaking *work*. Not vision, not tirades against the Evil Empire of Redmond, not manifestos. It needs a lot of people to sit down and clean things up. Small things, things that collectively annoy the user, but individually are too boring for an OSS developer to bother with.
That's how good software, how good *anything* gets done. The OSS way however, has been that for something to be refined, polished, it has to "scratch an itch", it has to annoy someone enough *and* have that someone be skilled or talented enough to go fix it him/herself.
How do you get people to work on minute trivial things that they normally donot care about? the one time-tested way is, well, to *pay* them to do it. Where do you get the money? well, you will probably need a corporate structure that will fund these developers and that will be able to stand on its own two feet.
In other words you need companies, companies like Active State, theKompany, Digital Creations (now Zope) and a few (very few) others. You need to let companies sell OSS without bitching all the time that they are ruining your free lunch. You need to let companies have pay-to-play versions that are ahead of the OSS one so that they can financially support development, QA and documentation. You need to have companies have non-GPL licenses on their products without going Homeini on them.
Preferrably all of the above do not involve flammage and mail-bombing and invocations of Rights and Freedom. Writing software is not Speech, it is *not* equivalent to expressing one's opinion on common matters ("politics", people call them). Writing software is hard, painful and mostly boring work, engineering work. It needs good design, a painstaking devotion to quality and most of all someone to be paying that budget.
... I was modeling combustion during my undergrad years as far back as '93, and there were already several labs across the country doing computational combustion work. For a variety of reasons computational combustion is one of the hardest simulation fields you can tackle. Here's a plug to the leading lab of my old prof, the Computational Combustion Laboratory at Georgia Tech
Well, that's not quite true; you can argue that a language newbie will pick the most 'natural' way to write something in a given language. If that's true (and I know that's arguable in itself), then a newbie's benchmark is a pretty good comparison of how the languages stack up.
E.g.: I am a Python freak, but for some things, 'regular' Python idioms (like for loops) are slow, but functional Python features (like map()) are much faster. I would use map() but that would be cheating, 'coz I wouldn't know the same shortcut for Perl, for example...
A speculative answer since b-trees are my bread and butter (I am just now specing a 2TB data-mine): hundreds of thousands of entries (or hundreds of millions) should not really bother a b-tree. From the articles about Google, I am guessing they have implemented some sort of distributed b-tree app server, across all those COTS linux boxes.
:-)...
I am curious as to what kind of implementation they are using; Google's roots would suggest some hacked form of Berkeley DB with lots of performance improvements.
Oh, well, just some guesswork... if I am close, I am expecting a job offer by the way
...Netscape will become Pathfinder for the GenX crowd. Only that when Pathfinder was live, only us "GenXers" (I hate that term too) were using the Web in the first place, and you know how Pathfinder turned out.
:-)...
The browser war is dead anyway. The coming war will be Web-based services (think IM, stock quotes, money management); MS has the better technology, but AOL has the users... it will be an interesting battle