If that worked, and Netscape could write code overnight, then in the morning you'd have a Netscape that was exactly like IE. Some people would use it because it's not M$oft, others wouldn't bother (they have IE, and what improvement would they get from another download). If Netscape didn't "embrace and extend" what IE did, then it would only ever be playing second fiddle to it.
Secondly (and this is the killer), using M$oft as a de facto standard is incredibly dangerous. M$oft have a demonstrated ability to pull huge U-turns when they want to, no matter what the cost. Imagine having a Netscape that had stopped tracking the W3C standards, tracked IE5 perfectly, then got entirely stomped by a sudden shift in IE6's behaviour.
I don't understand what right we all have to be blasting Microsoft over this
Agreed. If M$oft put an idiot-feature into IE, then that's great. Don't like it ? - Don't use it.
If you want to blast someone, blast the bozo NewMedia designers (curse their tiny little glasses and their pointy beards) who use such features without understanding the implications.
I love IE. It's stable (usually), powerful, and is the only useful XML client out there. Where the features it provides are generally helpful and likely to be (or become) mainstream, then I'll happily use them. I write SMIL that only works on IE 5.5, because it's my only option for SMIL, and in my particular context that's enough reason to change browsers. As SMIL is standards-based, then I have no guilt about doing it (Mozilla can play catch up as soon as they feel like it).
I'd love for there to be more good browsers. I'd love Mozilla to do XML (Yes, I know what it does, and that isn't useful enough). I'd love Amaya to be more friendly than a rottweiler with toothache. I'd like Opera to understand Unicode (big Doh! on that one, guys). These are business issues though, and as a web-geek, I'm not in a position to fix them. Hey, I'm just a red-shirt, and I know what happens when they go up against the Borg.
OTOH, M$oft "innovations" are evil, not part of the standards process, and should be shunned by all right thinking web developers. If M$oft want to use them on their own site, then that's their privilege and their problem if it goes wrong. No-one else should touch them with the proverbial bargepole.
PLEASE, browser makers - give us working, reliable CSS and a standard DOM before you fool around with anything else.
If you switch from DTDs to Schemas, then writing a similar tool becomes almost trivial, as an XSLT stylesheet. I use my own, tweaking it to suit the job in hand.
As a Schema is itself an XML document, unlike a DTD, it opens up all sorts of fun extensions to it like this. I haven't used DTDs in over a year, and I certainly don't miss them.
At JavaOne, Sun were also showing a similar product called "Adelard"
Yes, financial information needs to be kept. In the B&N case, this means that card was charged on in store . It doesn't require that the book titles themselves were logged though.
For marketing reasons, B&N will almost certainly do data warehousing and OLAP on this information. If people who buy Harry Potter also buy Ursula le Guin, then that's valuable information for them to know. OTOH, sensible OLAP often anonymises the data almost immediately to avoid many of these issues. IMHO, B&N having Lewinsky book title logs >6 months after the event is exactly the situation I'm saying they should try and avoid.
For the ISP, I still say that they should only log "core business" information, and be very careful about that. Maybe traffic logs to World-O-pr0n in the offshore Kinakuta data haven are a good indication that you need more bandwidth going out that way, but anything that attaches that traffic to other packets going to the billing system in Sealand is going to be majorly dodgy, from a customer privacy standpoint. Bookshops do their "traffic analysis" by looking at what individuals buy together, but ISPs only need to look at the aggregates (maybe tailored portals will be different).
Thanks for that, I'd mod it up if I had the points. Do you have a URL for some of this ? I'd like to see it.
2,000 getting T-1 run with a $250 dollar monthly fee. OTOH, that's not something I could do in the UK. Our on-line costs for something that can host servers are rather more than that. Our best entry point is rented Cobalts in a co-lo ($120/month) or US hosting.
It's sometimes said that programmers can only learn one language, then they spend the rest of their lives writing the same language, but with different syntax. Fortunately for me, my first language was Pascal. I might write Java today, but I write good Java, not some bastardised C++ (like so many ex-C people).
Pascal has been really good to me over the years. I'm in my mid-30s and my first freelance business was built on cheap clone PCs and Turbo Pascal -- like most of my geek generation.
That said, WHY in this day and age, is anyone still flogging the dead Pascal horse ? OK, so if this is compatible with Borland's OO extensions, then it's not a bad way of teaching OOP, but it's far from being the best. New tools are always cool and something to be encouraged, but please don't take this thing and use it on new projects, because you think Pascal is a good choice to be making (unless you're in a legacy support position).
The ISP should log nothing, out of their own self-interest. Anything they need to log, for their own purposes, should be destroyed after use.
Although it may be useful to Starr to find Lewinsky's book buying history, it's not good press for Barnes & Noble to have the existence of this log disclosed. Similarly it's never going to be in the ISPs interests to be at risk of having logs subpoenaed. The only legally-secure defence against this is to not have the logs in the first place (and this may require a traceable and provabel process to show that any that did exists have been destroyed).
UK IT industry on-line mag The Register has just run the SmartDownload story.
I'm glad the guys at Vulture Central ran it, because I certainly didn't get to read this story on the main/. page. Looking at the tiny number of comments so far, I wasn't the only one to miss it.
I write with a very expensive product for the day job, so I know it well, so it's what I use in the evenings too. If a Linux developer shells out money for Kylix, then they're damn well going to use the thing!
This doesn't mean The End Of Open Source (tm), but it is a risk. There will be code written under Kylix that gets OS'ed, and this will reduce the proportion of the total that's truly freely available.
I'm a Brit, so our "enclaves of poverty" aren't from quite the same source as yours, but we certainly have them.
You say you've managed to reduce unemployment and improve the local economy, by small-scale e-Commerce. Care to share any advice on how to do this ? I've worked on big over-funded startups that completely failed to make money, so I'm a little wary of suggesting eCommerce as a "guaranteed M M F" strategy.
What's the deal here ? What did you find worked, and what didn't ? What are the important skills to impart to people before they start the business ?
With the ability to handle such high traffic it would be suitable for busy e-commerce sites
eCommerce sites don't grind to a halt because they can't serve enough static pages, but because they can't grind the back-end processing fast enough(whether that be CGI Perl or Servlets).
OK, so some of them fall over because they've got their heads so far up their open sources that using a non-transactional database has become such a point of religious dogma that they won't go to any more sensible platform...
Is having the web server in the kernel a security hazard ? Does it increase the potential for damage if there is something like a buffer-overflow exploit in there too ? There are good reasons why the muggles and their IP-tomfoolery are normally kept out of the kernel.
Can you imagine the uproar if M$oft announced a new IIS embedded in kernel32.dll ?
The hardware differences somewhat devalued the original SPECWeb benchmarks.
If the pointy-haired stupids see this, they just say, "Oh but look, they slugged the results by running Linux on faster kit"
If the smarter people see it, they'll be kept too busy explain-o-LARTing to the stupids how irrelevant this is to be able to get full use of it in the Great Jihad against Redmond.
... anbd if M$oft's PR Weevils see it, they'll probably whine to the DoJ about unfair competition.
I've seen AUTOEXEC.BAT files with an ECHO Really Fast 486 at 50MHz in the first line (a few years back, when clock relabelling first started to be an issue for dodgy retailers). A few weeks ago I even saw a similar message in the boot.ini of an NT box ! Back in the days of Turbo buttons, some used to swear blind that the little LED displays on the front of the case (remember them ?) were an accurate measure of clock speed.
If you want to stop customers being ripped off by dubious retailers, you need to start with smarter, or better informed, customers. Telling them that clock speed doesn't matter(*) and that bundled WinModems are evil would be a good start. PC buyers are still naive and falling for bad deals that only look impressive in an advert.
(*) If it really matters to you, then it's your problem to find out how to check it. Don't be a label K1DD13 who thinks it's K00L just because it has the equivalent of a Tommy Hilfiger label on it.
"I don't care" is a good policy for dealing with privacy issues in a reasonably liberal Western democracy.
So what if Toysmart have your data ? It's just another piece of junkmail to deal with. Junkmail is like Quentin Crisp's technique for not hoovering the carpets; if you try to get every last dust-bunny, then you'll spend your life doing it. If you just let it build up, then after a while it doesn't get any thicker. Junk volumes are asymptotic - after a while they reach a maximum, you can just dump the lot en masse and get on with life.
Now before I have every Cryptognome in the subnet on my back, yes I'm aware how important these issues are. -- but this is a _toyshop_ FFS ! They're going to find how old my kids are and target me with "Save for College" plans - that's hardly house arrest in Lahore because a family member is a politician. Privacy issues are important, but a personal approach of just letting Toysmart have your data isn't unreasonable, and certainly not moronic (even if many morons also practice it).
I wish the FTC luck with this (and with anyone else who tries it). As a Lessigite (sic) privacy advocate, I believe in maintenance of presonal freedom by strong government WHERE APPROPRIATE, and this type of reining in of big business is where using government is appropriate.
Re:WAP not as popular as expected
on
WAP Under Fire
·
· Score: 1
In one sense, WAP is competing with GPRS.
Some months back, I investigated some Palm-based work with WAP and WML. What I discovered was twofold; that using WAP meant "the whole package" -- I couldn't run HTML over WAP, and I couldn't run WML over HTTP. If I wanted to host anything at all on the WAP mobile networks and gateways, I was very tightly constrained into my content.
Secondly (and more seriously), WML is useless for anything bigger than a phone; and by "bigger" I mean such huge devices as Palms and PDAs, not workstations. Issues like the WAP deck size limit are so painful that the protocol is simply unusable for anything above the most trivial of phone-like devices (6 line screens - runs out of steam for 12 line)
I also found that the much-discussed single XML format -> XSLT -> HTML + WML model of content production didn't work either. WML and its limits are so pervasive that they break the XML representation for your content, at a level far higher than a mere XSLT fix-up can deal with.
IMHO, WAP is dying here in the UK, and good riddance to it. It's another Rabbit Phone, so lets let it die and get one with GPRS and (hopefully) a decent content language like XHTML-Basic.
Sounds like a cunning idea, but with 100's of people an hour pouring through a portal, wouldn't cross contamination be an issue??
Cross contamination is already an issue, especially with drugs and explosives.
In the UK, we have the infamous case of "The Birmingham Six", supposed bombers of a pub in Birmingham in the mid '70s. They were convicted on the basis of slightly suspect forensic evidence linking them to having handled explosives recently, and some poor defence in court that didn't tear the credibility of this evidence to shreds. Chemically speaking, any nitrated polymer (paints, varnishes, glazes on shiny playing cards) is detectable by tests that look for nitrated polymers (certain sorts of explosive). Legally speaking, they were found guilty of Being Irish In A Public Place (a crime in England for much of the '70s). Fortunately this very case now serves as quite an effective warning against similar miscarriages of justice, as legislators are worried about perpetrating (or ratehr, becoming famous as the perpetrator of) a similar miscarriage.
As for drug residues, then just test a banknote. The proportion of these with cocaine traces is already some large fraction of those in circulation, especially in a large city. It's even being suggested as a convenient metric for studying drug usage in cities, the current drugs of choice, and some of their sourcing attributes (heroin is now only detectable, but you can tell roughly where it came from, the preparation method, and whether it was consumed as smoking or by injection). What you certainly can't do is track it to individual users.
This is a repressive and inappropriate use of technology and should be resisted by any means possible.
with joins involving more than seven tables, my experience has shown that MySQL just cannot optimize correctly,
Is a seven table join really a realistic case for MySQL's area of competence ?
I come from the "big iron" end of client-server databases, and the PHP / MySQL solutions are still pretty new to me. Now I can live without triggers, and I can code around a lack of foreign key constraints, but I'd start to feel a bit uneasy joining 7 tables on this grade of kit. For one thing; is it really necessary to build a join of that complexity (which will run slowly anywhere) on a low-budget web site ? Without transactions I'd be really concerned about being able to update that many tables and maintain reliable consistency anyway. Without stored procedures I'd be afraid of performing complex updates without a huge overhead at the connectivity level.
This whole bench test is pretty useless. Ziff wanted an "application" benchmark that was cross-platform and didn't rely on applications. What they actually built was so content-free that it simply tests network and OS performance as far as the TCP/IP stack.
Not surprisingly, they found that the Win TCP stack is quicker than the (known to be single-threaded) Linux stack. QEFD.
I'd like to see better benchmarks, but I'd much rather see something for simple Corba vs. Corba, or Corba vs DCOM. SOAP (the Apache approach of deployable handlers), vs. SOAP (Servlets) vs. SOAP (Microsft's SOAP-on-a-ROPE) would be even more interesting. We're doing something along those lines ourselves - maybe it will be publically publishable.
To get the alternative "Useless benchmark shows Linux to be faster than Windows" story go here.
Slashdot Mirror
cant netscape just comply with what IE does?
If that worked, and Netscape could write code overnight, then in the morning you'd have a Netscape that was exactly like IE. Some people would use it because it's not M$oft, others wouldn't bother (they have IE, and what improvement would they get from another download). If Netscape didn't "embrace and extend" what IE did, then it would only ever be playing second fiddle to it.
Secondly (and this is the killer), using M$oft as a de facto standard is incredibly dangerous. M$oft have a demonstrated ability to pull huge U-turns when they want to, no matter what the cost. Imagine having a Netscape that had stopped tracking the W3C standards, tracked IE5 perfectly, then got entirely stomped by a sudden shift in IE6's behaviour.
Macromedia, whose Flash animation technology is the de facto standard for lightweight vector graphics on the Web,
Flash ? Lightweight ? That'll be why boo was such a resounding success then.
I don't understand what right we all have to be blasting Microsoft over this
Agreed. If M$oft put an idiot-feature into IE, then that's great. Don't like it ? - Don't use it.
If you want to blast someone, blast the bozo NewMedia designers (curse their tiny little glasses and their pointy beards) who use such features without understanding the implications.
I love IE. It's stable (usually), powerful, and is the only useful XML client out there. Where the features it provides are generally helpful and likely to be (or become) mainstream, then I'll happily use them. I write SMIL that only works on IE 5.5, because it's my only option for SMIL, and in my particular context that's enough reason to change browsers. As SMIL is standards-based, then I have no guilt about doing it (Mozilla can play catch up as soon as they feel like it).
I'd love for there to be more good browsers. I'd love Mozilla to do XML (Yes, I know what it does, and that isn't useful enough). I'd love Amaya to be more friendly than a rottweiler with toothache. I'd like Opera to understand Unicode (big Doh! on that one, guys). These are business issues though, and as a web-geek, I'm not in a position to fix them. Hey, I'm just a red-shirt, and I know what happens when they go up against the Borg.
OTOH, M$oft "innovations" are evil, not part of the standards process, and should be shunned by all right thinking web developers. If M$oft want to use them on their own site, then that's their privilege and their problem if it goes wrong. No-one else should touch them with the proverbial bargepole.
PLEASE, browser makers - give us working, reliable CSS and a standard DOM before you fool around with anything else.
If you switch from DTDs to Schemas, then writing a similar tool becomes almost trivial, as an XSLT stylesheet. I use my own, tweaking it to suit the job in hand.
As a Schema is itself an XML document, unlike a DTD, it opens up all sorts of fun extensions to it like this. I haven't used DTDs in over a year, and I certainly don't miss them.
At JavaOne, Sun were also showing a similar product called "Adelard"
There =IS= no "new media" or "old media".
If there's no "New Meejah", why have my glasses got smaller, the sleeves fell off my fleece and I seem to have sprouted a little pointy goatee ?
Yes, financial information needs to be kept. In the B&N case, this means that card was charged on in store . It doesn't require that the book titles themselves were logged though.
For marketing reasons, B&N will almost certainly do data warehousing and OLAP on this information. If people who buy Harry Potter also buy Ursula le Guin, then that's valuable information for them to know. OTOH, sensible OLAP often anonymises the data almost immediately to avoid many of these issues. IMHO, B&N having Lewinsky book title logs >6 months after the event is exactly the situation I'm saying they should try and avoid.
For the ISP, I still say that they should only log "core business" information, and be very careful about that. Maybe traffic logs to World-O-pr0n in the offshore Kinakuta data haven are a good indication that you need more bandwidth going out that way, but anything that attaches that traffic to other packets going to the billing system in Sealand is going to be majorly dodgy, from a customer privacy standpoint. Bookshops do their "traffic analysis" by looking at what individuals buy together, but ISPs only need to look at the aggregates (maybe tailored portals will be different).
Thanks for that, I'd mod it up if I had the points.
Do you have a URL for some of this ? I'd like to see it.
2,000 getting T-1 run with a $250 dollar monthly fee.
OTOH, that's not something I could do in the UK. Our on-line costs for something that can host servers are rather more than that. Our best entry point is rented Cobalts in a co-lo ($120/month) or US hosting.
It's sometimes said that programmers can only learn one language, then they spend the rest of their lives writing the same language, but with different syntax. Fortunately for me, my first language was Pascal. I might write Java today, but I write good Java, not some bastardised C++ (like so many ex-C people).
Pascal has been really good to me over the years. I'm in my mid-30s and my first freelance business was built on cheap clone PCs and Turbo Pascal -- like most of my geek generation.
That said, WHY in this day and age, is anyone still flogging the dead Pascal horse ? OK, so if this is compatible with Borland's OO extensions, then it's not a bad way of teaching OOP, but it's far from being the best. New tools are always cool and something to be encouraged, but please don't take this thing and use it on new projects, because you think Pascal is a good choice to be making (unless you're in a legacy support position).
The ISP should log nothing, out of their own self-interest. Anything they need to log, for their own purposes, should be destroyed after use.
Although it may be useful to Starr to find Lewinsky's book buying history, it's not good press for Barnes & Noble to have the existence of this log disclosed. Similarly it's never going to be in the ISPs interests to be at risk of having logs subpoenaed. The only legally-secure defence against this is to not have the logs in the first place (and this may require a traceable and provabel process to show that any that did exists have been destroyed).
UK IT industry on-line mag The Register has just run the SmartDownload story.
I'm glad the guys at Vulture Central ran it, because I certainly didn't get to read this story on the main /. page. Looking at the tiny number of comments so far, I wasn't the only one to miss it.
Bad Slashdot. No VC cash.
Yes, very well written and clear - hope you're listening over at Ars Technica. 8-)
If only all technical explanations could be so well done (fancy explaining how to drive the IBM SOAP toolkit someday ?).
I write with a very expensive product for the day job, so I know it well, so it's what I use in the evenings too. If a Linux developer shells out money for Kylix, then they're damn well going to use the thing!
This doesn't mean The End Of Open Source (tm), but it is a risk. There will be code written under Kylix that gets OS'ed, and this will reduce the proportion of the total that's truly freely available.
The conference opened with a "Matrix" themed intro, and there have been blue and red jellybeans all over the place.
What happens if you eat the blue beans ?
Do you wake up in Redmond ?
I'm a Brit, so our "enclaves of poverty" aren't from quite the same source as yours, but we certainly have them.
You say you've managed to reduce unemployment and improve the local economy, by small-scale e-Commerce. Care to share any advice on how to do this ? I've worked on big over-funded startups that completely failed to make money, so I'm a little wary of suggesting eCommerce as a "guaranteed M M F" strategy.
What's the deal here ? What did you find worked, and what didn't ? What are the important skills to impart to people before they start the business ?
>Just about everyone here has more then 3 brain cells,
Is that really true on Slashdot these days ? Compare it now (beer-troll, etc.) to how it was two years ago. 8-(
> and know what [...] userland is.
These days they barely know their Dave Winers form their Dave Winders....
With the ability to handle such high traffic it would be suitable for busy e-commerce sites
eCommerce sites don't grind to a halt because they can't serve enough static pages, but because they can't grind the back-end processing fast enough(whether that be CGI Perl or Servlets).
OK, so some of them fall over because they've got their heads so far up their open sources that using a non-transactional database has become such a point of religious dogma that they won't go to any more sensible platform...
Firstly, I'm no Linux expert or OS expert.
Is having the web server in the kernel a security hazard ? Does it increase the potential for damage if there is something like a buffer-overflow exploit in there too ? There are good reasons why the muggles and their IP-tomfoolery are normally kept out of the kernel.
Can you imagine the uproar if M$oft announced a new IIS embedded in kernel32.dll ?
The hardware differences somewhat devalued the original SPECWeb benchmarks.
If the pointy-haired stupids see this, they just say, "Oh but look, they slugged the results by running Linux on faster kit"
If the smarter people see it, they'll be kept too busy explain-o-LARTing to the stupids how irrelevant this is to be able to get full use of it in the Great Jihad against Redmond.
... anbd if M$oft's PR Weevils see it, they'll probably whine to the DoJ about unfair competition.
I've seen AUTOEXEC.BAT files with an
ECHO Really Fast 486 at 50MHz
in the first line (a few years back, when clock relabelling first started to be an issue for dodgy retailers). A few weeks ago I even saw a similar message in the boot.ini of an NT box ! Back in the days of Turbo buttons, some used to swear blind that the little LED displays on the front of the case (remember them ?) were an accurate measure of clock speed.
If you want to stop customers being ripped off by dubious retailers, you need to start with smarter, or better informed, customers. Telling them that clock speed doesn't matter(*) and that bundled WinModems are evil would be a good start. PC buyers are still naive and falling for bad deals that only look impressive in an advert.
(*) If it really matters to you, then it's your problem to find out how to check it. Don't be a label K1DD13 who thinks it's K00L just because it has the equivalent of a Tommy Hilfiger label on it.
"I don't care" is a good policy for dealing with privacy issues in a reasonably liberal Western democracy.
So what if Toysmart have your data ? It's just another piece of junkmail to deal with. Junkmail is like Quentin Crisp's technique for not hoovering the carpets; if you try to get every last dust-bunny, then you'll spend your life doing it. If you just let it build up, then after a while it doesn't get any thicker. Junk volumes are asymptotic - after a while they reach a maximum, you can just dump the lot en masse and get on with life.
Now before I have every Cryptognome in the subnet on my back, yes I'm aware how important these issues are. -- but this is a _toyshop_ FFS ! They're going to find how old my kids are and target me with "Save for College" plans - that's hardly house arrest in Lahore because a family member is a politician. Privacy issues are important, but a personal approach of just letting Toysmart have your data isn't unreasonable, and certainly not moronic (even if many morons also practice it).
I wish the FTC luck with this (and with anyone else who tries it). As a Lessigite (sic) privacy advocate, I believe in maintenance of presonal freedom by strong government WHERE APPROPRIATE, and this type of reining in of big business is where using government is appropriate.
In one sense, WAP is competing with GPRS.
Some months back, I investigated some Palm-based work with WAP and WML. What I discovered was twofold; that using WAP meant "the whole package" -- I couldn't run HTML over WAP, and I couldn't run WML over HTTP. If I wanted to host anything at all on the WAP mobile networks and gateways, I was very tightly constrained into my content.
Secondly (and more seriously), WML is useless for anything bigger than a phone; and by "bigger" I mean such huge devices as Palms and PDAs, not workstations. Issues like the WAP deck size limit are so painful that the protocol is simply unusable for anything above the most trivial of phone-like devices (6 line screens - runs out of steam for 12 line)
I also found that the much-discussed single XML format -> XSLT -> HTML + WML model of content production didn't work either. WML and its limits are so pervasive that they break the XML representation for your content, at a level far higher than a mere XSLT fix-up can deal with.
IMHO, WAP is dying here in the UK, and good riddance to it. It's another Rabbit Phone, so lets let it die and get one with GPRS and (hopefully) a decent content language like XHTML-Basic.
Sounds like a cunning idea, but with 100's of people an hour pouring through a portal, wouldn't cross contamination be an issue??
Cross contamination is already an issue, especially with drugs and explosives.
In the UK, we have the infamous case of "The Birmingham Six", supposed bombers of a pub in Birmingham in the mid '70s. They were convicted on the basis of slightly suspect forensic evidence linking them to having handled explosives recently, and some poor defence in court that didn't tear the credibility of this evidence to shreds. Chemically speaking, any nitrated polymer (paints, varnishes, glazes on shiny playing cards) is detectable by tests that look for nitrated polymers (certain sorts of explosive). Legally speaking, they were found guilty of Being Irish In A Public Place (a crime in England for much of the '70s). Fortunately this very case now serves as quite an effective warning against similar miscarriages of justice, as legislators are worried about perpetrating (or ratehr, becoming famous as the perpetrator of) a similar miscarriage.
As for drug residues, then just test a banknote. The proportion of these with cocaine traces is already some large fraction of those in circulation, especially in a large city. It's even being suggested as a convenient metric for studying drug usage in cities, the current drugs of choice, and some of their sourcing attributes (heroin is now only detectable, but you can tell roughly where it came from, the preparation method, and whether it was consumed as smoking or by injection). What you certainly can't do is track it to individual users.
This is a repressive and inappropriate use of technology and should be resisted by any means possible.
with joins involving more than seven tables, my experience has shown that MySQL just cannot optimize correctly,
Is a seven table join really a realistic case for MySQL's area of competence ?
I come from the "big iron" end of client-server databases, and the PHP / MySQL solutions are still pretty new to me. Now I can live without triggers, and I can code around a lack of foreign key constraints, but I'd start to feel a bit uneasy joining 7 tables on this grade of kit. For one thing; is it really necessary to build a join of that complexity (which will run slowly anywhere) on a low-budget web site ? Without transactions I'd be really concerned about being able to update that many tables and maintain reliable consistency anyway. Without stored procedures I'd be afraid of performing complex updates without a huge overhead at the connectivity level.
This whole bench test is pretty useless. Ziff wanted an "application" benchmark that was cross-platform and didn't rely on applications. What they actually built was so content-free that it simply tests network and OS performance as far as the TCP/IP stack.
Not surprisingly, they found that the Win TCP stack is quicker than the (known to be single-threaded) Linux stack. QEFD.
I'd like to see better benchmarks, but I'd much rather see something for simple Corba vs. Corba, or Corba vs DCOM. SOAP (the Apache approach of deployable handlers), vs. SOAP (Servlets) vs. SOAP (Microsft's SOAP-on-a-ROPE) would be even more interesting. We're doing something along those lines ourselves - maybe it will be publically publishable.
To get the alternative "Useless benchmark shows Linux to be faster than Windows" story go here.