And in the.NET case they succeeded on a collosal scale.
And, tactics like this over the past decade and a half have damaged the software industry on a collosal scale. Consider a plausibly entertaining scenario: all the younguns are running around barking vaporware, business try to buy into, essentially, nothing, get confused why they aren't going anyware, a whole budget just got blown, managers are wondering what the hell is happening, saving face is no longer possible, a manager is fired and has only vaporware as his recent resume entry. This make for a pretty good tragic comedy.
Re:It takes insight to notice these things take ti
on
.Net:... 3 Years Later
·
· Score: 4, Funny
You can't really consider C# or Hailstorm to have been around and competing for three years, can you?
With Microsoft, yes, we can. Anyway, I was suprised to read that it's been three years already. This means we're due for.NET's replacement next year. Perhaps, this time around, we'll see a microkernel architecture with XXML (extensible XML, yea!) all implemented within a web services-based virtual machine. With that in 2004, I can't wait for 2008!
Exactly, so he and everybody else is sitting back and waiting for a clear winner with mature functionality to materialize.
Not necessarily. When put off by Hailstorm/Passport, Sun and dozens of other large private enterprises met and came up with the Liberty Alliance. Ostensibly, they came up with standards that allowed single-sign-on, even among businesses that don't trust each other.
What GM was probably tired of is new standards getting shoved not only down their throats but up their asses, too, by Microsoft and all the "me too" consultancy firms out there, who mainly employ doe-eyed college graduates and trendy-but-naive middle managers.
Note that at projectliberty.org, General Motors is a big member of the Libery Alliance, so they aren't entirly sitting on the side-lines, either.
Their business portal product is based on BizTalk, their.NET CRM application talks to the financial application through BizTalk, and they still have their e-commerce packages that are based on plain-old ASP and COM+.
This is one big problem with Microsoft. Each time some VP gets all horny for an idea, it seems whatever preceded that idea becomes somehow irrelevant from marketing and, eventually, support standpoints. I would bet there are many many millions of lines of commercial code out there tied by thier guts to COM, BizTalk, and whatever, leaving those Microsoft customers mystified about why they put forth all that effort only to have their vendor spit all over them and push them into the mud. Microsoft has absolutely no sense of being committed to their customers, which, IMO, is a big no-no in pretty much every other industry ever dreamt up by humans. This three-to-four year turnover in technology from them just needs to slow down (hell, the last major inventions elsewhere, UNIX, the WWW, Lisp/Java/VM-stuff etc. are all pretty darn "old", now, but evolving rather than getting uprooted and mulched).
Money for game setup? What does that even mean? There is no setup, you buy the game, put it in the console...
Buying the game and putting it in the console is pretty much what I meant by "set up". It seems many (most?) on-line games require purchasing something up-front before the recurring fees begin.
Monthly money for ISP service? How is that specific for online gaming? Since you're posting this on Slashdot, I assume you have some way of getting on the Internet as well.
The bread-n-butter of the average internet user is dial-up access via AOL, Juno, etc. at $8/month up to $22/month. If an on-line game works best with broad-band, then that would be an additional $20 to $40 per month. I suppose it depends on what each game requires.
First of all, the fee is annual and not monthly ($50 is pretty damn low for a year) and secondly that gets you an account on any Live game.
That is still a recurring fee, but $50/year is getting closer to a good amount (basically the cost of one game).
...the Xbox online pricing model.
Microsoft will find other ways to get their money from their users. The free marketing data they glean from Xbox Live might be enough, for now, until they learn enough or aquire enough to be able to pull their more aggressive tactics.
I wonder if he's dumb enough to think they'll sign a non-disclosure?
Given that he learned Japanese from Krusty the Klown ("Me so sowry!"), he will be thrown out of Japan and barred by the Japanese government from ever returning. I don't think we need to bother with speculation about NDAs.
On-line gameplay looks very expensive to me (money for network adapter, money for game set-up, monthly money for ISP service, monthly money for each on-line game). I always remembered Nintendo at its best with either single-player games or two or four-way games that kids would play with friends. I even played single-player games with friends, where we would take turns or play in a driver/navigator mode.
It seems that on-line games are still in their infancy and are probably fueled most by PC gamers who already incurred the cost of the computer and ISP service. To PC gamers the game fees are really only an incremental cost that is more easily tolerated.
For consoles to really catch on on-line, the prices really need to get driven down, because the main reason for sucessful consoles is large numbers of people too cheap to buy a $2500 gaming PC. For example, consider an average family who recently justified the cost of a cell phone and cable TV in the past several years now confronted with whether to shell out more money per month on on-line games. It took over a decade for cell phones to be in everyone's pocket and often displacing land-line service (rich and poor, it seems); perhaps it will be similar for on-line games.
I've used PCs now for decades, have Amigas and Sinclair QLs and, goddamnit, Dragon 32s still in working order, but I don't have a SINGLE fricking machine that relies upon a fan working that's more than three years old.
For PCs, this may be true, but for better hardware, such as (hopefully) the PowerMac G5, the fans should last at least five years, not three. This has been my experience with Sun workstations, where the ball-bearing case fans begin crapping out after five years of 24/7 operation (five years powered-on, not five years OS uptime, of course).
Actually, this is the correct way to write software. Reduce the number of lines. Increase the speed of execution.
No, it is not. Let's take Sed or Perl, for example. It is possible to do a temendous amount of work in one line of code using a complex Regular Expression, but debugging or maintaining that one line of code can be a nightmare, especially in shell scripts, where three or four parsers are operating on top of one another.
The same can occur in C regarding mixtures of structures, arrays, pointers, and function pointers. I can guarantee that the programmer will lose more hair debugging function pointers in structures or prefix increments within dereferences to arrays of structure pointers or whatever than they would lose had they originally written something readable.
Hard-to-read code is also more likely to get trashed, re-written, trashed, re-written, ad nauseum. Only people who write undocumented trash solely in the interest of job security would consider the-whole-world-in-two-lines a good programming model.
The truly correct way to write software is to write it first, then profile it to optimize the bottlenecks. Of course, the overall archtecture of the first draft must be reasonable, too, lest the whole program need optimizing.
I was quite impressed by the fact that Apache can cram all the functionality into ~59k lines.
Agreed. It would be interesting to know whether this low LOC is accomplished through good architecture that emphasizes simplicity and maintainability or "clever" hacks that compress a 10-line loop down into a three-line abomination of pointer arithmetic. I genuinely hope it is not the latter.
Regardless, 59K lines is small enough a program that--given a good architecture--can be studied and debugged relatively easily by one or two people. I'd estimate that this is why Apache is known for its low number of exploits in spite of its enormous web server market share.
Speaking from wishing I had rented Bounty Hunter and being glad I rented The Clone Wars, I wonder if Star Wars Galaxies is any good?
One thing I've noticed about the Star Wars franchise's games is that they take a genuinely interesting story and the outstanding Star Wars music and wrap them in a so-so game. I keep finding myself attracted to these games, but have resolved to steer away from them from now on. I don't know, perhaps these games are just dandy to teenagers--but, then, the violence in Bounty Hunter is stretching it for the teen market (e.g., killing hundreds of prison guards is just lovely).
For most professionals, there are no downsides for LCDs. Most jobs don't require super-accurate colors or ultra-small pixels (although Apple seems satisifed with LCDs), nor have I noticed problems viewing good LCDs from angles.
LCDs are appropriate for every non-gaming non-animation task I can think of, right now.
We rolled out IE5.01 using the IEAK (Internet Explorer Administration Kit). It would be a great thing if one could customize Mozilla in straight-foward manner for corporate deployments.
I don't know much about Mozilla on Windows, but Mozilla/Netscape 7 on Unix and Linux has everything stored in two directories, the install directory and the user's.mozilla directory.
If Mozilla is accessed via NFS, adding site-wide configurations should be trivial.
BTW, this is true of a network-installed StarOffice/OpenOffice.org, too. Again, NFS makes things very nice.
NFS, Mozilla, and OpenOffice.org licensing cost are, um....carry the zero...multiply by one..., uh, the licensing costs are zero!
If someone likes a particular program, and they know it well enough not to cause a load on your help desk, is there a real reason not to let them use it?
Good reasons to not use Outlook/Exchange:
1) What the hell is this MS-TNEF attachment crap? When I get one of these I have to use a separate program to decode it into a useful form.
2) Those graphical/HTML signatures are driving me crazy!
3) Bastardized non-MIME-multipart HTML messages are as stupid as the people who programmed Outlook in the first place! What's even more stupid is correct MIME-multipart messages that send the body of the message twice! (My point is that HTML in e-mail is stupid however it is done)
4) Outlook is the single reason why my company uses a draconian SPAM filter that generates an embarrasing number of false positives. Oh, an entire server is deditcated to virus-scanning every attachment recieved or sent.
5) When you really look at it, Outlook isn't even a good productivity application. Sure it does stuff, but it is only used by everyone because it came with Office. I remember the organizer packaged with SmartSuite as being much better. Emacs+vm is a more functionally-complete e-mail client. Netscape has a clearer user interface.
In conclusion, Outlook is like a cheap hooker who does lots of things, but leaves the customers feeling unsure about their hygine and whether their moral convictions were worth what just happened.
Re:I don't have an answer, but...
on
RAID for Zero-G?
·
· Score: 1
And you'd do that for > 500GB? That's more than 5,000 microdrives... I don't think that's more efficient than a couple of 200GB drives.
Actually, it would be 500 microdrives; regardless, I agree that would be too many.
Fifteen 80GB laptop drives, however, would be very practical, but they might end up weighing more than seven or so 200GB drives. I guess it all depends on the final numbers (storage to weight ratio).
The chapter reminds us that guaranteeing support for 6 operating systems and 4 database backends and 3 browsers means that we have to support and provide quality assurance for 6x4x3=72 combinations of products.
I believe very strongly that portable coding is possible and practical. The fact that Visual Basic is so alluring to the lazy should be no excuse. There are such things as Java, POSIX, ANSI SQL, ANSI C, etc.. Most frequently, deviations from these standards are small and add functionality, such as BLOBS in SQL, that aren't consistently implemented, yet. These deviations can be supported by isolating them in the software and providing abstractions that make them invisible to the rest of the application. This is called good architecture. I'm sorry that there are so many people out there who are too stubborn, lazy, and/or stupid to recognize the benefits of good architecture and portability.
The cost analyses that "prove" that non-portable software are better most likely include false assumptions about the cost of supporting additional platforms. They usually leave out the costs saved by organizing the software well, which makes support cheaper through fast problem resolution, fast support for new requirements, etc. Addionally, what are the costs of rewriting from scratch when the chosen platform becomes obselete or the vendor tanks? I'd say those costs are so great that creating portable software should be the rule rather than the exception.
For example, how many companies would simply go bankrupt if Microsoft went they way of Enron? I'd say that our economy is much more fragile than most people will admit.
For example, Microsoft's "marketecture" is actually a lying-through-the-teeth-marketing department coupled with an unscrupulous aquire-and-crush department coupled with leadership that has a psychological deformity that makes them believe in world domination. Their "tarchitecture" are too-smart-for-their-own-good college graduates and a counter-intuitive culture of cut-n-paste and stock options.
In the "real world", which Microsoft is not a part of, there is no distinction between "marketecture" and "tarchitecure". Well designed software is genuinely useful, with a healthy set of features, a healthy user interface, and robustness. I believe I will call it, simply, "architecture." If this "architecture" does not truly fill the customers need, then capitalistic natural selection provides an honest solution.
You may claim that Microsoft has truly filled the needs of many people, but, upon closer inspection, this is not true. Many others have found ways to fill the same needs better, but, somehow, Microsoft skirted the laws of nature and invented their own microuniverse of delusion. I believe they will be short-lived in the grand scheme of things, where truly superior architectures will come in and make them obselete.
Re:I don't have an answer, but...
on
RAID for Zero-G?
·
· Score: 2, Funny
Shuttle middeck.
You know, with as rediculously expensive as it is per kilogram to launch stuff, you might want to rig a custom enclosure that uses laptop hard drives or microdrives (those little 1GB postage-stamp things). They have IDE interfaces, so I'm sure one rigging the controllers wouldn't be too hard, and you could use software RAID from one of those miniature Via-CPU motherboards.
As far as G-forces go, pack it all in bubblewrap, which would be entertaining for the crew, as well.
The heat is a concern, and the lower the power dissipation, the better.
I was suprised how cool, quiet, and low-vibration the 10KRPM SCSI drives I have run (latest Seagate and a fairly new Fujitsu). Perhaps simply researching among the newest models of drives will give you adequate power consumption/heat generation.
The right enclosure is pretty important, too. For example, one nice thing about the high-end Sun workstations (Ultra 60, Blade 2000) is that the disks are mounted with generous space between them and direct air flow from big case fans. There is practically no chance of overheating in these enclosures.
What the hell is wrong with you? This kind of coverage only causes trouble.
What is wrong with you? This coverage has spoiled the event for all practical purposes. For example, I forwarded the SFGate article to our webmaster.
So, while you are so pessimistic, please realize that, now, millions of sysadmins can go ahead and double-check their patchlevels, etc.
I don't mean to sound mean, but full disclosure allows natural selection to run its course more efficiently among software companies and consumers alike. It also continually brings us closer to a day, when "software engineering" isn't an oxymoron and a joke.
If people really knew what was inside that Windows 9x or 200x shrinkwrap, they would probably explode with a sharp popping sound, after their brains suddenly realize they had been lied to for over a decade.
The USPS has changed from competition from the likes of FedEx and UPS, and they are now very, very good at what they do.
I was suprised the other day to be able to get a certified mail tracking number from the USPS. It allows going to usps.com and verifying the package was signed for, and its only a few dollars. That's another way the USPS is a good alternative to UPS and FedEX.
The letter at grc.com constantaly spells "sniffer" as "SNIFFER(R)". How can the two be confused? I've heard "sniffer" for years without any mention of Computer Associates. If I saw "SNIFFER(R)", then there's no confusion that it is probably a product of some kind, but confusing "sniffer" simply doesn't make sense (it's a generic term for software that sniffs--it's a verb, too!--packets from a network transmission).
By the way, from the letter: "This includes, but is not limited to, the use of "Sniffer" in any meta tags, source code, key words, domain names, glossaries, indexes and the like associated with your web site(s)."
Slashdot Mirror
And in the .NET case they succeeded on a collosal scale.
And, tactics like this over the past decade and a half have damaged the software industry on a collosal scale. Consider a plausibly entertaining scenario: all the younguns are running around barking vaporware, business try to buy into, essentially, nothing, get confused why they aren't going anyware, a whole budget just got blown, managers are wondering what the hell is happening, saving face is no longer possible, a manager is fired and has only vaporware as his recent resume entry. This make for a pretty good tragic comedy.
You can't really consider C# or Hailstorm to have been around and competing for three years, can you?
.NET's replacement next year. Perhaps, this time around, we'll see a microkernel architecture with XXML (extensible XML, yea!) all implemented within a web services-based virtual machine. With that in 2004, I can't wait for 2008!
With Microsoft, yes, we can. Anyway, I was suprised to read that it's been three years already. This means we're due for
Exactly, so he and everybody else is sitting back and waiting for a clear winner with mature functionality to materialize.
Not necessarily. When put off by Hailstorm/Passport, Sun and dozens of other large private enterprises met and came up with the Liberty Alliance. Ostensibly, they came up with standards that allowed single-sign-on, even among businesses that don't trust each other.
What GM was probably tired of is new standards getting shoved not only down their throats but up their asses, too, by Microsoft and all the "me too" consultancy firms out there, who mainly employ doe-eyed college graduates and trendy-but-naive middle managers.
Note that at projectliberty.org, General Motors is a big member of the Libery Alliance, so they aren't entirly sitting on the side-lines, either.
Their business portal product is based on BizTalk, their .NET CRM application talks to the financial application through BizTalk, and they still have their e-commerce packages that are based on plain-old ASP and COM+.
This is one big problem with Microsoft. Each time some VP gets all horny for an idea, it seems whatever preceded that idea becomes somehow irrelevant from marketing and, eventually, support standpoints. I would bet there are many many millions of lines of commercial code out there tied by thier guts to COM, BizTalk, and whatever, leaving those Microsoft customers mystified about why they put forth all that effort only to have their vendor spit all over them and push them into the mud. Microsoft has absolutely no sense of being committed to their customers, which, IMO, is a big no-no in pretty much every other industry ever dreamt up by humans. This three-to-four year turnover in technology from them just needs to slow down (hell, the last major inventions elsewhere, UNIX, the WWW, Lisp/Java/VM-stuff etc. are all pretty darn "old", now, but evolving rather than getting uprooted and mulched).
Money for game setup? What does that even mean? There is no setup, you buy the game, put it in the console...
...the Xbox online pricing model.
Buying the game and putting it in the console is pretty much what I meant by "set up". It seems many (most?) on-line games require purchasing something up-front before the recurring fees begin.
Monthly money for ISP service? How is that specific for online gaming? Since you're posting this on Slashdot, I assume you have some way of getting on the Internet as well.
The bread-n-butter of the average internet user is dial-up access via AOL, Juno, etc. at $8/month up to $22/month. If an on-line game works best with broad-band, then that would be an additional $20 to $40 per month. I suppose it depends on what each game requires.
First of all, the fee is annual and not monthly ($50 is pretty damn low for a year) and secondly that gets you an account on any Live game.
That is still a recurring fee, but $50/year is getting closer to a good amount (basically the cost of one game).
Microsoft will find other ways to get their money from their users. The free marketing data they glean from Xbox Live might be enough, for now, until they learn enough or aquire enough to be able to pull their more aggressive tactics.
I wonder if he's dumb enough to think they'll sign a non-disclosure?
Given that he learned Japanese from Krusty the Klown ("Me so sowry!"), he will be thrown out of Japan and barred by the Japanese government from ever returning. I don't think we need to bother with speculation about NDAs.
On-line gameplay looks very expensive to me (money for network adapter, money for game set-up, monthly money for ISP service, monthly money for each on-line game). I always remembered Nintendo at its best with either single-player games or two or four-way games that kids would play with friends. I even played single-player games with friends, where we would take turns or play in a driver/navigator mode.
It seems that on-line games are still in their infancy and are probably fueled most by PC gamers who already incurred the cost of the computer and ISP service. To PC gamers the game fees are really only an incremental cost that is more easily tolerated.
For consoles to really catch on on-line, the prices really need to get driven down, because the main reason for sucessful consoles is large numbers of people too cheap to buy a $2500 gaming PC. For example, consider an average family who recently justified the cost of a cell phone and cable TV in the past several years now confronted with whether to shell out more money per month on on-line games. It took over a decade for cell phones to be in everyone's pocket and often displacing land-line service (rich and poor, it seems); perhaps it will be similar for on-line games.
I've used PCs now for decades, have Amigas and Sinclair QLs and, goddamnit, Dragon 32s still in working order, but I don't have a SINGLE fricking machine that relies upon a fan working that's more than three years old.
For PCs, this may be true, but for better hardware, such as (hopefully) the PowerMac G5, the fans should last at least five years, not three. This has been my experience with Sun workstations, where the ball-bearing case fans begin crapping out after five years of 24/7 operation (five years powered-on, not five years OS uptime, of course).
Actually, this is the correct way to write software. Reduce the number of lines. Increase the speed of execution.
No, it is not. Let's take Sed or Perl, for example. It is possible to do a temendous amount of work in one line of code using a complex Regular Expression, but debugging or maintaining that one line of code can be a nightmare, especially in shell scripts, where three or four parsers are operating on top of one another.
The same can occur in C regarding mixtures of structures, arrays, pointers, and function pointers. I can guarantee that the programmer will lose more hair debugging function pointers in structures or prefix increments within dereferences to arrays of structure pointers or whatever than they would lose had they originally written something readable.
Hard-to-read code is also more likely to get trashed, re-written, trashed, re-written, ad nauseum. Only people who write undocumented trash solely in the interest of job security would consider the-whole-world-in-two-lines a good programming model.
The truly correct way to write software is to write it first, then profile it to optimize the bottlenecks. Of course, the overall archtecture of the first draft must be reasonable, too, lest the whole program need optimizing.
I mean, after it's compiled, why do you care?
Because, I'm the one doing the compiling!
I was quite impressed by the fact that Apache can cram all the functionality into ~59k lines.
Agreed. It would be interesting to know whether this low LOC is accomplished through good architecture that emphasizes simplicity and maintainability or "clever" hacks that compress a 10-line loop down into a three-line abomination of pointer arithmetic. I genuinely hope it is not the latter.
Regardless, 59K lines is small enough a program that--given a good architecture--can be studied and debugged relatively easily by one or two people. I'd estimate that this is why Apache is known for its low number of exploits in spite of its enormous web server market share.
Speaking from wishing I had rented Bounty Hunter and being glad I rented The Clone Wars, I wonder if Star Wars Galaxies is any good?
One thing I've noticed about the Star Wars franchise's games is that they take a genuinely interesting story and the outstanding Star Wars music and wrap them in a so-so game. I keep finding myself attracted to these games, but have resolved to steer away from them from now on. I don't know, perhaps these games are just dandy to teenagers--but, then, the violence in Bounty Hunter is stretching it for the teen market (e.g., killing hundreds of prison guards is just lovely).
Crack whores cost more than fatties!
You can get a better value by requesting the Assistant Crack Whore-in-Training. They haven't mastered the trade, yet, but they get the job done.
Yeah, I can't print in color but that does not bother me.
Laser printers are perfect for legal documents and resumes, which are about all that's worth printing, anymore.
The only uses I've seen for color inkjets are home-made birthday cards and rather unprofessional-looking digital photographs.
BTW, has anyone been brave enough to do a color resume, yet?
...they don't know that the LCDs have downfalls.
For most professionals, there are no downsides for LCDs. Most jobs don't require super-accurate colors or ultra-small pixels (although Apple seems satisifed with LCDs), nor have I noticed problems viewing good LCDs from angles.
LCDs are appropriate for every non-gaming non-animation task I can think of, right now.
the antitrust service pack
For those of you who haven't heard of this, yet, here are places you can get it:
Mirror 1a
Mirror 1b
Mirror 2
Mirror 3
Mirror 4
These mirrors also work well:
Mirror 5
Mirror 6
We rolled out IE5.01 using the IEAK (Internet Explorer Administration Kit). It would be a great thing if one could customize Mozilla in straight-foward manner for corporate deployments.
.mozilla directory.
I don't know much about Mozilla on Windows, but Mozilla/Netscape 7 on Unix and Linux has everything stored in two directories, the install directory and the user's
If Mozilla is accessed via NFS, adding site-wide configurations should be trivial.
BTW, this is true of a network-installed StarOffice/OpenOffice.org, too. Again, NFS makes things very nice.
NFS, Mozilla, and OpenOffice.org licensing cost are, um....carry the zero...multiply by one..., uh, the licensing costs are zero!
Why are you using Windows and IE, again???
If someone likes a particular program, and they know it well enough not to cause a load on your help desk, is there a real reason not to let them use it?
Good reasons to not use Outlook/Exchange:
1) What the hell is this MS-TNEF attachment crap? When I get one of these I have to use a separate program to decode it into a useful form.
2) Those graphical/HTML signatures are driving me crazy!
3) Bastardized non-MIME-multipart HTML messages are as stupid as the people who programmed Outlook in the first place! What's even more stupid is correct MIME-multipart messages that send the body of the message twice! (My point is that HTML in e-mail is stupid however it is done)
4) Outlook is the single reason why my company uses a draconian SPAM filter that generates an embarrasing number of false positives. Oh, an entire server is deditcated to virus-scanning every attachment recieved or sent.
5) When you really look at it, Outlook isn't even a good productivity application. Sure it does stuff, but it is only used by everyone because it came with Office. I remember the organizer packaged with SmartSuite as being much better. Emacs+vm is a more functionally-complete e-mail client. Netscape has a clearer user interface.
In conclusion, Outlook is like a cheap hooker who does lots of things, but leaves the customers feeling unsure about their hygine and whether their moral convictions were worth what just happened.
And you'd do that for > 500GB? That's more than 5,000 microdrives... I don't think that's more efficient than a couple of 200GB drives.
Actually, it would be 500 microdrives; regardless, I agree that would be too many.
Fifteen 80GB laptop drives, however, would be very practical, but they might end up weighing more than seven or so 200GB drives. I guess it all depends on the final numbers (storage to weight ratio).
The chapter reminds us that guaranteeing support for 6 operating systems and 4 database backends and 3 browsers means that we have to support and provide quality assurance for 6x4x3=72 combinations of products.
I believe very strongly that portable coding is possible and practical. The fact that Visual Basic is so alluring to the lazy should be no excuse. There are such things as Java, POSIX, ANSI SQL, ANSI C, etc.. Most frequently, deviations from these standards are small and add functionality, such as BLOBS in SQL, that aren't consistently implemented, yet. These deviations can be supported by isolating them in the software and providing abstractions that make them invisible to the rest of the application. This is called good architecture. I'm sorry that there are so many people out there who are too stubborn, lazy, and/or stupid to recognize the benefits of good architecture and portability.
The cost analyses that "prove" that non-portable software are better most likely include false assumptions about the cost of supporting additional platforms. They usually leave out the costs saved by organizing the software well, which makes support cheaper through fast problem resolution, fast support for new requirements, etc. Addionally, what are the costs of rewriting from scratch when the chosen platform becomes obselete or the vendor tanks? I'd say those costs are so great that creating portable software should be the rule rather than the exception.
For example, how many companies would simply go bankrupt if Microsoft went they way of Enron? I'd say that our economy is much more fragile than most people will admit.
Smoke and mirrors. How stupid.
For example, Microsoft's "marketecture" is actually a lying-through-the-teeth-marketing department coupled with an unscrupulous aquire-and-crush department coupled with leadership that has a psychological deformity that makes them believe in world domination. Their "tarchitecture" are too-smart-for-their-own-good college graduates and a counter-intuitive culture of cut-n-paste and stock options.
In the "real world", which Microsoft is not a part of, there is no distinction between "marketecture" and "tarchitecure". Well designed software is genuinely useful, with a healthy set of features, a healthy user interface, and robustness. I believe I will call it, simply, "architecture." If this "architecture" does not truly fill the customers need, then capitalistic natural selection provides an honest solution.
You may claim that Microsoft has truly filled the needs of many people, but, upon closer inspection, this is not true. Many others have found ways to fill the same needs better, but, somehow, Microsoft skirted the laws of nature and invented their own microuniverse of delusion. I believe they will be short-lived in the grand scheme of things, where truly superior architectures will come in and make them obselete.
Shuttle middeck.
You know, with as rediculously expensive as it is per kilogram to launch stuff, you might want to rig a custom enclosure that uses laptop hard drives or microdrives (those little 1GB postage-stamp things). They have IDE interfaces, so I'm sure one rigging the controllers wouldn't be too hard, and you could use software RAID from one of those miniature Via-CPU motherboards.
As far as G-forces go, pack it all in bubblewrap, which would be entertaining for the crew, as well.
The heat is a concern, and the lower the power dissipation, the better.
I was suprised how cool, quiet, and low-vibration the 10KRPM SCSI drives I have run (latest Seagate and a fairly new Fujitsu). Perhaps simply researching among the newest models of drives will give you adequate power consumption/heat generation.
The right enclosure is pretty important, too. For example, one nice thing about the high-end Sun workstations (Ultra 60, Blade 2000) is that the disks are mounted with generous space between them and direct air flow from big case fans. There is practically no chance of overheating in these enclosures.
What the hell is wrong with you? This kind of coverage only causes trouble.
What is wrong with you? This coverage has spoiled the event for all practical purposes. For example, I forwarded the SFGate article to our webmaster.
So, while you are so pessimistic, please realize that, now, millions of sysadmins can go ahead and double-check their patchlevels, etc.
I don't mean to sound mean, but full disclosure allows natural selection to run its course more efficiently among software companies and consumers alike. It also continually brings us closer to a day, when "software engineering" isn't an oxymoron and a joke.
If people really knew what was inside that Windows 9x or 200x shrinkwrap, they would probably explode with a sharp popping sound, after their brains suddenly realize they had been lied to for over a decade.
The USPS has changed from competition from the likes of FedEx and UPS, and they are now very, very good at what they do.
I was suprised the other day to be able to get a certified mail tracking number from the USPS. It allows going to usps.com and verifying the package was signed for, and its only a few dollars. That's another way the USPS is a good alternative to UPS and FedEX.
The letter at grc.com constantaly spells "sniffer" as "SNIFFER(R)". How can the two be confused? I've heard "sniffer" for years without any mention of Computer Associates. If I saw "SNIFFER(R)", then there's no confusion that it is probably a product of some kind, but confusing "sniffer" simply doesn't make sense (it's a generic term for software that sniffs--it's a verb, too!--packets from a network transmission).
By the way, from the letter: "This includes, but is not limited to, the use of "Sniffer" in any meta tags, source code, key words, domain names, glossaries, indexes and the like associated with your web site(s)."
This is simply assinine. Source code?