"I heard a statistic once that if you chose answers randomly on a MC test that you could get a C by not knowing anything beyond how to circle a letter!"
You "heard a statistic once"? Geez, the probability statistics aren't that difficult: If there's 4 possible answers, and you randomly pick, you'll likely get about 25% right, or 5/20, 3/33. It isn't rocket science. To get 50% randomly there'd have to be only two possible choices. Add to that the fact that many post secondary multiple choice tests actually deduct marks for incorrect answers, and your C proclamation sounds like it might be incorrect.
"As for mainframes - I just got done working on a box with 72 tetrabytes of RAM, shared over three states. It is not the same thing. Your right, I bet the itanium is a faster CPU than the slice of the system I got in my VM."
I'm rather curious what the specs are on the machine that you're working on. Would you mind given some details on this machine that gives you such incredible authority and insight (in your dorm?). Let me put it into perspective for you a bit: The machine I was talking about, on which the.NET platform runs, has 64 Itanium2 processors (though I'm sure you'll assure yourself that the Itanium2 pales compared to whatever you use...have some benchmarks to prove it?). For comparions, the IBM zSeries 990 (which is I believe their most powerful non-supercomputer "mainframe". The sort of things that only the hugest enterprises buy) has up to 32 less powerful processors. In other words, if you've missed the point, the J2EE scalability FUD is absolute trash: On the HP Integrity Superdome you can scale up to an exponential more powerful machine than probably 99.999% of Slashdot users, and J2EE fanatics, have ever dreaming of coming near. Yeah there are a few research computers faster, but they're not really the point when we're talking about scalability in a "real world" context.
"Odds are, if it could get by running it on a laptop - you don't need the J2EE framework. Tomcat and Struts are probably enough. "
Hilarious. Here's the funny thing: Many operations have at their core AS/400s that are absolutely pummelled by two-way Xeon machines you can purchase online from Dell. Very seldom nowadays, even for huge databases, is there a CPU issue if you have an adequate SAN. Where there is a CPU issue most modern designs (such as Google) rely on scaling out rather than buying a supercomputer regardless (it's more scalable, and generally more cost effective).
Of course, I guess this all just means that I'm not hanging with the big boys like you. Let's see: Nationwide life insurance system. Check. Enterprise-wide CRM system. Check. Yeah, I'll have to find some bigger gigs to put it into perspective for myself. Maybe I should use some less efficient tools to necessitate a surplus of computational power.
Re:I think you're confused.
on
Java vs .NET
·
· Score: 1
The SQL Server example, as I'm sure is clear to anyone without a religious war to wage (you tried implementing in.NET first huh....suurrrrreeeee you did), is a documented, easily accessible metric of a "Win32" (actually Win64) machine at mainframe types of scaling.
"Since microsoft only handles 2-4 CPU x86 boxes (in practice), that is one of the real limiting factors IMHO."
Firstly, this is classic "just in case" argumentation that people make in defense of Java, despite the fact that about, oh, 99% of applications won't need anything more than a 2-way server (and btw for those that are even remotely considering the "big unix boxes", 8 way Windows 2000 boxes are cheap and easy to come by). Of those applications that do require more, for about 3 years now "big iron" has been retro, and clustering has been the way to go (see Google).
Having said that, you should reconsider your arguments. The #2 result on the TPC non-clustered results is SQL Server 2000 64-bit edition running on Windows 2003 64-bit edition. The server is a 64-way Itanium2 box (for those of you keeping count, this is far more powerful than the vast majority of "mainframes"). Microsoft implementation of.NET for the 64-bit platform is still in beta, however existing.NET assemblies will run flawlessly on it instantly.
I'm saying nothing about whether.NET or Java is better, but am just saying that the "Java scales,.NET doesn't" BS is nothing but pure unadulterated FUD.
"Enough of these rubbish do-dads that frankly, no good driver would ever need."
Traction control and anti-lock brakes both can accomplish feats that the best driver on the planet Earth couldn't accomplish, given that they are modulating their input thousands of times per second, absolutely optimizing power/braking and acceleration. They aren't necessities, and you could approximate them, but don't fool yourself into thinking that you could do a better job. If you claim otherwise, then I'd suggest that you should have relay wires installed in your dash to let you handle the spark timing yourself as well.
Where technology has just finally offered a better choice is in transmissions: Until recently the "standard" was always the superior driving choice, as the automatic options were hydraulic circuitry turds often with 3 gears. Now there are continuously variable automatic transmissions that achieve the absolute perfect coupling between power plant and road surface all of the time. Alternately the computer controlled 5 or 6 speed automatics are quite extraordinary now as well.
Errr...you basically just repeated exactly what he was overtly implying. When he mentioned "hacking", clearly he was talking about overriding the limiter.
This is also the America where consumers can ignore all of the informationpummelledinto them, make poor consumer choices, but then amazingly they can turn around and profess a child-like ignorance, actually suing because they should be protected from their own poor judgement.
Quality and security of software is a market feature, and if the public ignores the continual security lapses of some particularly popular software, for instance, and if they accept that there will be X crashes per week, then so be it: The marketplace has spoken. We don't need anyone protecting us from ourselves, and feigning ignorance after the fact is incredibly weak.
"...an exploit for a vulnerability patched a month in advance..."
For a hole that was in the system for years, which is similar to many other major in the news exploits. The fact that the patch was available for months is little consolation if there were nefarious groups who were aware of these holes for years, which is something that no one can conclusively answer.
I think the simplistic "all other systems are secure, but MS systems are weak" zealotry often repeated by the puppets is incredibly weak, but at the same time let's face the fact that there are likely hundreds (or thousands) more exploits on every Windows machine out there, silently waiting to be exploited. (Linux may have as many or more, but I'm not talking about that here). It disturbs me to think that there are very likely countries and groups doing the same research that companies like eEye do, but perhaps they don't have a business model that relies upon publishing exploits for media PR...instead they keep them under their belts for selective and intelligent use when necessary (rather than the Ebola like high school student worm).
Perhaps the month long security audit at Microsoft was a good step forward, however there is no doubt that it will be a massive undertaking to basically give the entire codebase an enema, removing ridiculously trivial exploits like buffer overflows. The security issues in Microsoft code is much more than a month long effort: Microsoft must put a massive, concerted, effective effort at securing their code, because each time another buffer overflow exploit comes out, or an exploit for a trivial service that absolutely no one uses (internet printer service, home automation plug & play), it makes them look like a completely amateur shop that can't be trusted.
Firstly this is just yet another doom-and-gloom BS article of the sort that appears during every single downturn: Each time it's pronouncement of prophecies, and then a few years later when we have a market where web slugs are making $150K/annum these people are silently biding their time waiting for the next downturn to spout their negativity.
Having said that, firstly Indian workers aren't working for "less" : Many of them have large homes, servents, etc. The issue is one of currency conversion: The US dollar is grossly overvalued, and while it allows US companies to buy foreign firms cheap, it also makes the same US operations uncompetitive on the global market (which is why the US has had a trade deficit for many years). Already as the US $ has declined the hypothetical cost competitiveness of Indian firms has greatly diminished.
In the end, though, India isn't the "problem" with the IT market: The problem is that IT hasn't delivered on its promise. In many organizations the redundant and overlapping IT processes take a large share of the budget, earning a lot of attention for cost savings. The software development process is an absolute FARCE, with the majority of software projects being absolute failures, often coupled with extremely heavyweight processes that ensure that the actual developer is a tiny portion of the process (with a massive business paper trail). Tell me that you can get a 30% savings by outsourcing to India, and I'd say that you could probably yield a 80%+ savings by culling the deadweight and switching to an Agile process: Something that actually yields results.
So who do you blame for that? Sounds more like a Dell problem, in that apparently they're loading crippleware and time expired software on their systems and offering it as a feature.
My stereo has a knob marked "volume". Why would record companies (or artists, or whomever we are blaming) record something "louder" on a CD?
You're exactly right: This article might be a bit misleading -- It has nothing whatsoever to do with loudness, as loudness is up to the player/amp. It also doesn't have anything to do with "lack of headroom" (because they aren't dumb enough to actually clip on a digital medium because that sounds like unbelievable garbage). The problem is that it's very common nowadays to turn an entire song into what I guess one could call "mono-level": They compress it, from a dynamic range perspective, so that instead of a soft passage being 12db less than a loud passage, it's 3db. They do this because, quite honestly, the average consumer demands it -- People get upset if they have to turn their stereo up in the car to hear the soft sections, and then get shocked when a loud chorus comes on.
I'll second that about the fluid bearing Western Digital: Extremely quiet. I have the power saving mode that spools the HD down, a habit from prior configurations where the hard drive was the loud part and it was some aureal relief, however with this drive it's imperceptible if it's actually spinning or not.
The Volcano 7 heatsink on the other hand...what a mistake.
They really are. Sites like the NYT that require a free subscription often do so to provide a reader profile to their advertisers, and it is used to show the ads to the appropriate people. i.e. Ads for goods and services only available in the US aren't giving value to the advertiser when viewers from around the globe are seeing them, for instance.
So true. I'm talking in Canadian dollars here, but I remember the slick little maneuver, one which the media happily played along with (claiming that the auto industry, which amazingly pulled through the "recession" generally making hefty profits, was "losing money" on their "interest free" deals), being exemplified in the Chrysler Neon: Suddenly a $14K CDN car was selling for $20K at 0% interest, and people were lining up to buy them. 0% indeed.
Sidenote related to your post, but Plastic has an excellent posting system, offering both a superb spell checker, but also a nice reader level style meter. Just an FYI about a very nice board system.
If most of their current operation is Windows based, wouldn't it make even more sense to equip all of their Windows boxes with VMWare and run Linux in a virtual session for the new, forward-moving apps, eventually switching over to a native Linux install when appropriate?
It sounds like this is completely a political and/or religious choice rather than a practical choice. Now each desktop has the licensing costs of Windows, Linux and VMWare. _S_M_R_T...
Re:Cassette Adapters
on
Pods Unite
·
· Score: 2, Insightful
but it would probably spur the people making portable devices to include optical outputs on them
Is there something inherently wrong with coaxial optical outputs? A lot of people seem to get a chubby about using optical, yet in this case, and even for multi-channel high-sample rate sound, a simple RCA coaxial cable is more than sufficient, and the quality is 100% (i.e. 100% of the bits make it through unscathed). Somehow thousands of suckers are out spending $40 for a 12" optical cable...must be right after they put the orange marker around their CDs.
The Adobe SVG viewer recently, and surprizingly, had a beta of version 6.0 come out. This can be found here. Of course because the Mozilla folks changed the plug-in mechanisms, it still crashes Mozilla, but this is great news from Adobe as many rumors were that the entire SVG team was canned. Obviously that isn't true, and they continue to develop upon it. Corel, as you obviously know if you read even the summary of this article, has a great viewer out as well, along with a variety of tools for working with SVG.
Of course MSDN Magazine covered SVGs last month. Of course I'm biased given that I wrote that article.
This whole ask Slashdot is just pure rhetoric tripe. C# triggers? Sounds like someone does something one way in one environment, so they expect to do it that way in every environment. The whole question about "scaling" is ignorant in any case: His question is really "How much can I do with X box", not a scaling question (which would be "How big of a box, or how many boxes in a cluster").
Here would be my answers, though this chap is so clueless it's barely worth it.
1. Says you. On the one hand this question is about bargain basement equipment and setup, but it needs to be a high load messaging server? Uh huh.
2. This question should be "Should I use SOAP?"..NET in no way ties one to SOAP, nor does SOAP tie you to.NET (use SOAP on Apache).
3. "I have a hammer so why can't every problem be a nail? I don't understand RDBMS'...I'm confused!"
4. What sort of claptrap BS is this? The SQL Server numbers are not only best for the highest TPC, but also for most TPCs for the cheapest amount. Pure BS question.
5. Uh...
6. Uh...okay...don't use flipping reflection then. Absolutely ridiculous.
7. Firstly this is pure, unadulterated BULLSHIT (ASPX absolutely SMOKES ASP in virtually all situations), but even ignoring that, what does this have to do with.NET? WebServices is one solution offered in the.NET grab bag, just like you can do it in Apache.
8. Completely ignorant.
9. I will somewhat give this. Replication on SQL Server, like on most database servers, is a bitch, however it works great when done correctly. However I suspect that someone is just digging in a bag of concepts because for what they're talking about it would be wholly inappropriate.
How this pure bullshit got posted on the front page boggles the mind. Actually, no, I just realized that this is Slashdot...it doesn't boggle the mind at all. In fact it makes total sense.
Next up : "Does Windows 2000 Scale?"
"I used Windows 2000 on a Pentium 60 with 128MB of RAM, and when I call WaitForSingleEventEx sometimes it takes more than 60ms!"
I don't normally write short agreement posts, but your post is brilliant. Very well put.
Of course I will take issue with one thing: There are a lot of managers out there who refuse to listen to programmers who petition that the "One Universal Cross-Platform Middleware Library That Everyone Will Use Forever" isn't practically doable, and who'll simply rotate through the ranks until they get someone who'd nod their head and say yes at all the right times regarding the Object-Oriented Infinitely Configurable Endlessly Moldable Last-Software Product You'll Ever Need.
Ebay doesn't charge for the commission unless the deal settles. If "someone" bids a million dollars, and you report afterwards that they're a non-payer, there is no commission charged.
Seriously, though, to make it even worse about -- One second I'm sqeezing out a statistic [grunt] -- 99% of "newsworthy" auctions are outright frauds, whether it's for a town or for Eminem's childhood home: There are virtually no identification checks on Ebay for bidders, so I can go and auction off my Slashdot UID, auction myself up to a million dollars, and run around talking about the intrinsic value of Slashdot UIDs. Of course then The Turd Report (god bless his soul) will come along and set new records for UID value.
Slashdot Mirror
My organization uses a Linksys BEFSR41 for firewalling. Works great.
"I heard a statistic once that if you chose answers randomly on a MC test that you could get a C by not knowing anything beyond how to circle a letter!"
You "heard a statistic once"? Geez, the probability statistics aren't that difficult: If there's 4 possible answers, and you randomly pick, you'll likely get about 25% right, or 5/20, 3/33. It isn't rocket science. To get 50% randomly there'd have to be only two possible choices. Add to that the fact that many post secondary multiple choice tests actually deduct marks for incorrect answers, and your C proclamation sounds like it might be incorrect.
"As for mainframes - I just got done working on a box with 72 tetrabytes of RAM, shared over three states. It is not the same thing. Your right, I bet the itanium is a faster CPU than the slice of the system I got in my VM."
.NET platform runs, has 64 Itanium2 processors (though I'm sure you'll assure yourself that the Itanium2 pales compared to whatever you use...have some benchmarks to prove it?). For comparions, the IBM zSeries 990 (which is I believe their most powerful non-supercomputer "mainframe". The sort of things that only the hugest enterprises buy) has up to 32 less powerful processors. In other words, if you've missed the point, the J2EE scalability FUD is absolute trash: On the HP Integrity Superdome you can scale up to an exponential more powerful machine than probably 99.999% of Slashdot users, and J2EE fanatics, have ever dreaming of coming near. Yeah there are a few research computers faster, but they're not really the point when we're talking about scalability in a "real world" context.
I'm rather curious what the specs are on the machine that you're working on. Would you mind given some details on this machine that gives you such incredible authority and insight (in your dorm?). Let me put it into perspective for you a bit: The machine I was talking about, on which the
"Odds are, if it could get by running it on a laptop - you don't need the J2EE framework. Tomcat and Struts are probably enough. "
Hilarious. Here's the funny thing: Many operations have at their core AS/400s that are absolutely pummelled by two-way Xeon machines you can purchase online from Dell. Very seldom nowadays, even for huge databases, is there a CPU issue if you have an adequate SAN. Where there is a CPU issue most modern designs (such as Google) rely on scaling out rather than buying a supercomputer regardless (it's more scalable, and generally more cost effective).
Of course, I guess this all just means that I'm not hanging with the big boys like you. Let's see: Nationwide life insurance system. Check. Enterprise-wide CRM system. Check. Yeah, I'll have to find some bigger gigs to put it into perspective for myself. Maybe I should use some less efficient tools to necessitate a surplus of computational power.
The SQL Server example, as I'm sure is clear to anyone without a religious war to wage (you tried implementing in .NET first huh....suurrrrreeeee you did), is a documented, easily accessible metric of a "Win32" (actually Win64) machine at mainframe types of scaling.
"Since microsoft only handles 2-4 CPU x86 boxes (in practice), that is one of the real limiting factors IMHO."
.NET for the 64-bit platform is still in beta, however existing .NET assemblies will run flawlessly on it instantly.
.NET or Java is better, but am just saying that the "Java scales, .NET doesn't" BS is nothing but pure unadulterated FUD.
Firstly, this is classic "just in case" argumentation that people make in defense of Java, despite the fact that about, oh, 99% of applications won't need anything more than a 2-way server (and btw for those that are even remotely considering the "big unix boxes", 8 way Windows 2000 boxes are cheap and easy to come by). Of those applications that do require more, for about 3 years now "big iron" has been retro, and clustering has been the way to go (see Google).
Having said that, you should reconsider your arguments. The #2 result on the TPC non-clustered results is SQL Server 2000 64-bit edition running on Windows 2003 64-bit edition. The server is a 64-way Itanium2 box (for those of you keeping count, this is far more powerful than the vast majority of "mainframes"). Microsoft implementation of
I'm saying nothing about whether
"Enough of these rubbish do-dads that frankly, no good driver would ever need."
Traction control and anti-lock brakes both can accomplish feats that the best driver on the planet Earth couldn't accomplish, given that they are modulating their input thousands of times per second, absolutely optimizing power/braking and acceleration. They aren't necessities, and you could approximate them, but don't fool yourself into thinking that you could do a better job. If you claim otherwise, then I'd suggest that you should have relay wires installed in your dash to let you handle the spark timing yourself as well.
Where technology has just finally offered a better choice is in transmissions: Until recently the "standard" was always the superior driving choice, as the automatic options were hydraulic circuitry turds often with 3 gears. Now there are continuously variable automatic transmissions that achieve the absolute perfect coupling between power plant and road surface all of the time. Alternately the computer controlled 5 or 6 speed automatics are quite extraordinary now as well.
Errr...you basically just repeated exactly what he was overtly implying. When he mentioned "hacking", clearly he was talking about overriding the limiter.
This is also the America where consumers can ignore all of the information pummelled into them, make poor consumer choices, but then amazingly they can turn around and profess a child-like ignorance, actually suing because they should be protected from their own poor judgement.
Quality and security of software is a market feature, and if the public ignores the continual security lapses of some particularly popular software, for instance, and if they accept that there will be X crashes per week, then so be it: The marketplace has spoken. We don't need anyone protecting us from ourselves, and feigning ignorance after the fact is incredibly weak.
"...an exploit for a vulnerability patched a month in advance..."
For a hole that was in the system for years, which is similar to many other major in the news exploits. The fact that the patch was available for months is little consolation if there were nefarious groups who were aware of these holes for years, which is something that no one can conclusively answer.
I think the simplistic "all other systems are secure, but MS systems are weak" zealotry often repeated by the puppets is incredibly weak, but at the same time let's face the fact that there are likely hundreds (or thousands) more exploits on every Windows machine out there, silently waiting to be exploited. (Linux may have as many or more, but I'm not talking about that here). It disturbs me to think that there are very likely countries and groups doing the same research that companies like eEye do, but perhaps they don't have a business model that relies upon publishing exploits for media PR...instead they keep them under their belts for selective and intelligent use when necessary (rather than the Ebola like high school student worm).
Perhaps the month long security audit at Microsoft was a good step forward, however there is no doubt that it will be a massive undertaking to basically give the entire codebase an enema, removing ridiculously trivial exploits like buffer overflows. The security issues in Microsoft code is much more than a month long effort: Microsoft must put a massive, concerted, effective effort at securing their code, because each time another buffer overflow exploit comes out, or an exploit for a trivial service that absolutely no one uses (internet printer service, home automation plug & play), it makes them look like a completely amateur shop that can't be trusted.
Firstly this is just yet another doom-and-gloom BS article of the sort that appears during every single downturn: Each time it's pronouncement of prophecies, and then a few years later when we have a market where web slugs are making $150K/annum these people are silently biding their time waiting for the next downturn to spout their negativity.
Having said that, firstly Indian workers aren't working for "less" : Many of them have large homes, servents, etc. The issue is one of currency conversion: The US dollar is grossly overvalued, and while it allows US companies to buy foreign firms cheap, it also makes the same US operations uncompetitive on the global market (which is why the US has had a trade deficit for many years). Already as the US $ has declined the hypothetical cost competitiveness of Indian firms has greatly diminished.
In the end, though, India isn't the "problem" with the IT market: The problem is that IT hasn't delivered on its promise. In many organizations the redundant and overlapping IT processes take a large share of the budget, earning a lot of attention for cost savings. The software development process is an absolute FARCE, with the majority of software projects being absolute failures, often coupled with extremely heavyweight processes that ensure that the actual developer is a tiny portion of the process (with a massive business paper trail). Tell me that you can get a 30% savings by outsourcing to India, and I'd say that you could probably yield a 80%+ savings by culling the deadweight and switching to an Agile process: Something that actually yields results.
So who do you blame for that? Sounds more like a Dell problem, in that apparently they're loading crippleware and time expired software on their systems and offering it as a feature.
My stereo has a knob marked "volume". Why would record companies (or artists, or whomever we are blaming) record something "louder" on a CD?
You're exactly right: This article might be a bit misleading -- It has nothing whatsoever to do with loudness, as loudness is up to the player/amp. It also doesn't have anything to do with "lack of headroom" (because they aren't dumb enough to actually clip on a digital medium because that sounds like unbelievable garbage). The problem is that it's very common nowadays to turn an entire song into what I guess one could call "mono-level": They compress it, from a dynamic range perspective, so that instead of a soft passage being 12db less than a loud passage, it's 3db. They do this because, quite honestly, the average consumer demands it -- People get upset if they have to turn their stereo up in the car to hear the soft sections, and then get shocked when a loud chorus comes on.
More like a cross between C++ and Object Pascal in Delphi. Seriously. That anyone could hold Java up for originality boggles the mind.
I'll second that about the fluid bearing Western Digital: Extremely quiet. I have the power saving mode that spools the HD down, a habit from prior configurations where the hard drive was the loud part and it was some aureal relief, however with this drive it's imperceptible if it's actually spinning or not.
The Volcano 7 heatsink on the other hand...what a mistake.
They really are. Sites like the NYT that require a free subscription often do so to provide a reader profile to their advertisers, and it is used to show the ads to the appropriate people. i.e. Ads for goods and services only available in the US aren't giving value to the advertiser when viewers from around the globe are seeing them, for instance.
Of course MS does provide the .NET SDK for free, which allows one to create web applications, or even Windows forms applications.
So true. I'm talking in Canadian dollars here, but I remember the slick little maneuver, one which the media happily played along with (claiming that the auto industry, which amazingly pulled through the "recession" generally making hefty profits, was "losing money" on their "interest free" deals), being exemplified in the Chrysler Neon: Suddenly a $14K CDN car was selling for $20K at 0% interest, and people were lining up to buy them. 0% indeed.
Sidenote related to your post, but Plastic has an excellent posting system, offering both a superb spell checker, but also a nice reader level style meter. Just an FYI about a very nice board system.
If most of their current operation is Windows based, wouldn't it make even more sense to equip all of their Windows boxes with VMWare and run Linux in a virtual session for the new, forward-moving apps, eventually switching over to a native Linux install when appropriate?
It sounds like this is completely a political and/or religious choice rather than a practical choice. Now each desktop has the licensing costs of Windows, Linux and VMWare. _S_M_R_T...
but it would probably spur the people making portable devices to include optical outputs on them
Is there something inherently wrong with coaxial optical outputs? A lot of people seem to get a chubby about using optical, yet in this case, and even for multi-channel high-sample rate sound, a simple RCA coaxial cable is more than sufficient, and the quality is 100% (i.e. 100% of the bits make it through unscathed). Somehow thousands of suckers are out spending $40 for a 12" optical cable...must be right after they put the orange marker around their CDs.
The Adobe SVG viewer recently, and surprizingly, had a beta of version 6.0 come out. This can be found here. Of course because the Mozilla folks changed the plug-in mechanisms, it still crashes Mozilla, but this is great news from Adobe as many rumors were that the entire SVG team was canned. Obviously that isn't true, and they continue to develop upon it. Corel, as you obviously know if you read even the summary of this article, has a great viewer out as well, along with a variety of tools for working with SVG.
Of course MSDN Magazine covered SVGs last month. Of course I'm biased given that I wrote that article.
This whole ask Slashdot is just pure rhetoric tripe. C# triggers? Sounds like someone does something one way in one environment, so they expect to do it that way in every environment. The whole question about "scaling" is ignorant in any case: His question is really "How much can I do with X box", not a scaling question (which would be "How big of a box, or how many boxes in a cluster").
.NET in no way ties one to SOAP, nor does SOAP tie you to .NET (use SOAP on Apache).
.NET? WebServices is one solution offered in the .NET grab bag, just like you can do it in Apache.
Here would be my answers, though this chap is so clueless it's barely worth it.
1. Says you. On the one hand this question is about bargain basement equipment and setup, but it needs to be a high load messaging server? Uh huh.
2. This question should be "Should I use SOAP?".
3. "I have a hammer so why can't every problem be a nail? I don't understand RDBMS'...I'm confused!"
4. What sort of claptrap BS is this? The SQL Server numbers are not only best for the highest TPC, but also for most TPCs for the cheapest amount. Pure BS question.
5. Uh...
6. Uh...okay...don't use flipping reflection then. Absolutely ridiculous.
7. Firstly this is pure, unadulterated BULLSHIT (ASPX absolutely SMOKES ASP in virtually all situations), but even ignoring that, what does this have to do with
8. Completely ignorant.
9. I will somewhat give this. Replication on SQL Server, like on most database servers, is a bitch, however it works great when done correctly. However I suspect that someone is just digging in a bag of concepts because for what they're talking about it would be wholly inappropriate.
How this pure bullshit got posted on the front page boggles the mind. Actually, no, I just realized that this is Slashdot...it doesn't boggle the mind at all. In fact it makes total sense.
Next up : "Does Windows 2000 Scale?"
"I used Windows 2000 on a Pentium 60 with 128MB of RAM, and when I call WaitForSingleEventEx sometimes it takes more than 60ms!"
I don't normally write short agreement posts, but your post is brilliant. Very well put.
Of course I will take issue with one thing: There are a lot of managers out there who refuse to listen to programmers who petition that the "One Universal Cross-Platform Middleware Library That Everyone Will Use Forever" isn't practically doable, and who'll simply rotate through the ranks until they get someone who'd nod their head and say yes at all the right times regarding the Object-Oriented Infinitely Configurable Endlessly Moldable Last-Software Product You'll Ever Need.
Ebay doesn't charge for the commission unless the deal settles. If "someone" bids a million dollars, and you report afterwards that they're a non-payer, there is no commission charged.
Seriously, though, to make it even worse about -- One second I'm sqeezing out a statistic [grunt] -- 99% of "newsworthy" auctions are outright frauds, whether it's for a town or for Eminem's childhood home: There are virtually no identification checks on Ebay for bidders, so I can go and auction off my Slashdot UID, auction myself up to a million dollars, and run around talking about the intrinsic value of Slashdot UIDs. Of course then The Turd Report (god bless his soul) will come along and set new records for UID value.