Although I use C++ professionally myself, I'm not ready to say it's easy to debug or understand. The C++ templates, while powerfull, are difficult to debug when the projects get large. It can be hard to determine what they expand to. My debugger, in particular, gets confused with them. VSlick gets confused with the preprocessor tags, too.
Second point: I suspect the choice of language might be a Hobson's choice. Either they were told to use it, or must use it for performance considerations.
On Java: You don't fully win here either, because although it may be more portable, it isn't as stable. I don't mean realiable, I mean that Sun keeps on changing it much more frequently. It's also a very vast platform--just look at how many books at Barnes and Noble there are on the subject. Way too much to read.
Compiler: If C++ is going to be used, your choice of compiler could be as important as the library you use. If someone could suggest a good one for Windows that compiles to native windows with native libraries (not cygwin), that's free, I'd be more than interested.
I too wonder this. Having only used RedHat 7.x,
I can't say what the differences are between
distributions. I thought the end result was the same, once you got past the installations of the OS and applications. Not flamebait, but I'm just curious.
I haven't mastered Swing or AWT. I didn't want to spend time reading 100 pages on ListModels and such just to put up a stinking tree control. Is it any simpler in the SWT?
My gripe is, that I really miss GUI programming in VB, where things were nice and easy and it didn't require reading 500 pages.
What advanced storage techniques? If there are such, they need to be more prominently published. Also, if XML needs to be stored in a more compressed form, why can't XML be passed in a compressed form? Why can't it just be text compiled into a binary encoded format? Use commonly available open source code to read it.
Let's get away from this obsurd notion that everything has to be done in text. It's not faster. It's not more portable. Spelling tags out instead of using symbols doesn't make things platform independent. Just because you can't use a regular text editor doesn't make it harder to use. Just faster.
XML is another attempt by bad software engineering practices to slow down good hardware engineering practices.
Okay, so the marketers have a point. But how about at least having a plain option. Low frames. Low or no images. Or putting a description on images which act as buttons or such.
Remember, the more you embed your information in fancy web stuff, the harder it is to find it in a web search.
The idea that displaying text, excuse me, text with layout or markup tags, being slower than molases is not impressive. Does anyone else cringe when they have to open up a help link web link from a local application?
It takes really long because browser people insist on having a JRE, javascript interpreter, sometimes mail and newsreaders. Of course then there's the plugin's.
Whatever happened to the good ole days where text was text, and help files where man pages or a Windows help file. And desktop publishing was done on WordPerfect?
Yes. Yes. Yes. Do get angry at these web people. I used to be able to dial directly into my bank and download my transactions, and pay bills, all without a web browser. And it was faster. I don't care what you web people say. Life is faster when you don't spell everything out in plain text and use pretty graphics and javascript and such.
Yes. Get rid of the excessive javascript, or even better, don't use it at all! Get rid of the excessive pictures. Don't put a back picture when I could use my back key! Don't create popup menus, just use links. Don't put up ads on bank account pages, especially after the customer has paid you $6.95 per month.
And give the information! Don't make us email you for it. Don't make us call some 800 number and talk to a salesperson. If you have prices, put them up! Don't hide them unless you're ashamed of them.
Have honest links. If you have a download link for an application, for instance, don't make us go through 10,000 slow, image laden web pages just to download the thing. A download link should take us to a downloadable file! (Or a page with the OS selection and such). Forget the mirrors crap. Just ask us a location and direct us to it.
To the web developers: Make life simpler, and faster. Not slow and annoying!
I disagree about Flash. I really wish web developers would have the courtesy of not using things like this. The web protocol and most browsers with them, is really slow. It's also not innovative except in allowing people to pass whole words in the form of tags when they could pass symbols and save bandwidth. We don't need to make it any slower. So if using a standard such as Mosaic 1.0 saves bandwidth by cutting out the fancy crap, I'm all for it. I don't use the web for pretty pictures. I use it for research, and people who insist on developing software for the the absolute slowest GUI available.
I find the opposite to be true, except for Postgresql.
Unix=Windows-(Ease of Use+Decent OfficeApplications+Hardware drivers and support+Programmers who aren't cockier than their code deserves+Internet Explorer which face it, is still really good, inspite of you MS haters)
Windows isn't going away, and won't be supplanted by a Unix variant, Linux or otherwise, anytime soon.
Very few large companies have just one platform, and most have established applications on multiple platforms. Switching is cost prohibitive.
*nix is not revolutionary, it's just as much a legacy as mainframes.
The idea of everything related to system and software configuration being stored in slow to parse text, in multiple formats is Unix. Having a single registry, stored without wasting space, is revolutionary.
Further, Windows problems can usually be fixed by commonly available utilities, requiring little skill. Oh, I'm sorry, I forgot that I should be a kernel expert to maintain a *nix system.
*nix programs are poorly named, either cute or short, causing much confusion. Task Scheduler vs crond or anacronda. Of course we all know about the thousands of text readers/displayers, etc, such as nroff, xv, latex (I know the case is wrong).
Support and software availability is better on the Windows platform. OK, yeah I can get StarOffice or OpenOffice for Linux. But I can also get that on Windows, plus MS Office which is better than anything on Unix systems, plus WordPerfect (built for Windows and emulated on *nix systems), and Lotus SmartSuite.
All of the above is a different story of course, if you're comparing the two platforms for purely server use. In that case you would probably pick *nix, but not necessarily.
If PC's are fairly cheap, then why not just buy a 2nd PC and put Windows on it. You're going to have to buy it anyway. Might as well have the real thing.
I guess your humility may be a good start. Ego is way to plentiful in this business.
I saw a book on what it means to be a Software Architect, beyond drawing pretty diagrams, but also some political stuff, too. Look at Borders, sorry I don't remember anything else.
How prevalent is this role now? My company has them, but how likely would we be to see them in others?
There's two aspects to depreciation, however. From an accounting/tax perspective, accelerating depreciation is preferable, since it reduces taxable income faster. There is a school of thought that says depreciation is almost a fiction from a cash perspective, since the expense isn't a cash expense (at least not until the next purchase).
People need to stop associating the meaning of life with web browsing. Web browsing should be about looking at pages displayed clearly, and FAST. It shouldn't be about messaging, email, news groups, chicken soup and making coffee.
Netscape 4.x can't display pages clearly because whoever is responsible for X windows makes it a challenge to change the fonts used with it. Styles are window manager issues, not browser issues. Probably most of the interface problems are X problems, which though thoroughly worth griping about, aren't relevant. The last I tested Netscape 6, I think it had the same problems.
Fortunately for me, I've been using IE, which seems to be more stable then the above. It really helps to be using the hedgemonic browser. I'm not sure how these people got their IE browsers to crash, but I certainly remember Netscape crashing perpetually on Linux, since I had to muck with the X and Gnome/KDE settings to get it to work. I never have to do this with Windows, because It Just Works! And faster at that!
The reviewer did point out a few common needs every browser should have though:
It's nice to use a legible, clean and intelligently laid out web page. But flashy graphics, macromedia flashplayer crap, no thank you. Maybe all web developers should use telephone lines--see how they like all those frames, graphics only buttons, etc. Oh, and stop using that Adobe Acrocrap stuff. Just use HTML like the rest of the page.
A Few Points, which are almost never mentioned, because they're blasphemous to most web users. But first, I realize ASN.1 is a syntax specifying language used for intellegent information passing, unlike XML/XHTML/HTML which is for bloated, inefficient message passing. However, the, for me, point is speed, not syntax.
1) Its about time someone realized that spelling everything out into words and wordlike symbols actually costs something. Admit it. It takes longer to pass delimiters taking the form of then to come up with a byte encoded form of transmission.
2) Words are not more portable. I can understand a binary bitstream on an intel platform the same as on some Sun machine and even on an IBM mainframe using EBCDIC as long as I knew what it's codepage was. I don't need things spelled out to transfer it to a different hardware or OS platform. Try it sometime if you don't believe me.
3) Wordlike symbols are harder to translate. Its easier to attach multiple words, all meaning the same thing in different languages, to one byte encoded symbol than to an English like word.
4) Parsing speed. This is relevant because some people think XML is a good distributed message passing scheme between two machines--read machines, not human beings. It also doesn't require extra, exotic, hard to find, libraries either.
Oh, and stop saying computers are getting faster as though that were a justification. These poor hardware engineers are having a tough time keeping up with incompetent software engineers. With that attitude, the only reason computing might get faster is because of the hardware, certainly not the software. These people who say this need to go back to a 80386 so they can learn to program with resource restrictions--which produces good programs on faster machines.
5) Who cares if a human can read it. It takes special XML readers to format documents in the intelligent way they're meant to be. And those readers are really hard to find if you're cheap like me. Just try to do a search for "Free XML readers". You'll come up with thousands, most of which are just irritating shareware, or slow Java based programs. Further, humans can read all forms of communication, binary or otherwise, just use a debugger if neeeded. Most text editors have a hex viewing/editing capability. Yes, you'll have to do some thinking, and it will be slower to decipher, but it will be much SMALLER.
6) You don't need XML/HTML/XHTML or like markup languages to have a standard. You could standardize on a binary encoding. Yes, maybe we could have a lab experiment on this. I think we really could! Again, try to get people to agree on a syntax derived from human language. It's easier to pick an octet stream and let people decide what it means in their language.
I don't thoroughly hate XML/HTML/XHTML, or like ideas. People just misuse and over use them. It
Amen. Amen. Amen.
Let me begin my rant!
This problem isn't limited to Java though.
How many times hasn't that frustrated all of us. It's these irritating web jerks who have to make us surf through 80 billion links which all say "download now" just to finally get to the page. Plus, most sites usually want an abundance of information to get the "free" download, and another goddamn password. Then they have to give me a laundry list of download sites because no browser maker has yet to automate the process of finding a reliable site from which to download.
Also, we don't all have T1 lines going into our houses. Cable modems in my area cost $40+/month. Five megs for a JRE alone IS huge, and it would be nice for MS to provide this for that reason alone. I'm sorry if I don't enjoy spending 3 hours babysitting a modem connection. This all leads to my conspriracy theory that all software makers try their absolute hardest to slow down computers to make sure that they always run at the speed of a 286. Why else would we have things like Java, XML and the web?
Actually I used to work on a mainframe myself. OS/390 v2.4-2.8. And yes, I have used JCL rather extensively, and have found it do what you say--but not much more. It's a very precise way of executing files, and crude file management. Compare that with shell scripts which don't require column alignments and a// in column 1. Its advantage is simplicity in reading once you understand it, because you can't do much with it.
OS/390 is a really good OS internally. I just don't like everything being limited to units of 8 characters (PDS member names, portions of a DSN,etc). You also can't beat it for programs like RACF for power. With the Unix System Services (which includes vi!), it's gotten a lot more flexible and enables easier porting of UNIX programs. However, I've done admin work on them, from setting up RACF classes, users, TCP/IP stacks, some SMP/E, parmlib setups, etc. Reliable? Definately. Powerful? Usually. Easy? Almost never. Overall though, the architecture and design would make most people here on Slashdot feel in awe of it. Incidently, most of these complaints don't apply to VM, which is what Linux first ran on. I think I saw an article saying IBM got 40,000 partitions running.
I stopped working on them because my product didn't want any more Unix type work done, not because of hatred. I respect them quite a great deal.
I can sympathize with coworkers who pawn off work. Unfortunately, they don't need to be drunk to do this.
If a person was drunk you'd think that would show up in there prior experience, and perhaps the drug test. Sure they can sober up for an interview or a drug test, but how are you going to find that out in a credit check?
Secondly, we already tell private businesses how they must conduct hiring practices. In particular, how they may not conduct it. I'll argue that it's good to have anti-discrimination legislation. I've talked with hiring managers who say it's illegal to ask a company listed as a reference for more than dates of employment, and salary/wages. This is to prevent blacklisting. So if you can't get that information, which would be much more directly relevant to a hiring decision, why should you get the highly indirect/speculative and personal financial background of a candidate?
Further, companies and organizations have a right to keep trade secrets, and such, but why privacy? Companies and organizations are by nature far more exposed to the public than people are. For instance, financial statements are public in the case of corporations, and some companies. In short, private businesses aren't as private as you may think, but people are, and need protection.
I agree that this increases prestige for Linux. Mainframes have gotten an extremely bad reputation for several reasons that Linux would address.
1. Ease of use: Mainframes have historically been hard to use and/or interface with. Anyone who has worked with MVS or what is now known as OS/390 and next Z/OS would remember horrible things. Examples are JCL (Satan's version of a batch file or shell script), Record based file systems such as PDS's, VSAM datasets, and the like which either can't or are hard to directly copy or interface with, and must be preallocated before use. No GUI interface, and limited command line capability in ISPF. There is an X Windows capability via the UNIX services layer, but applications and utilities haven't been ported there much.
2. Culture--This is good and bad. Mainframe programmers tend to be very disciplined. They can also be resistent to change (ie. stubborn), slow because they're disciplined or because they're just slow.
Yet these are just the OS and cultural problems. This is what mainframe architecture gives and has given almost since conception.
1. Reliability.
If your DASD(permanent storage), CPU or other device goes out, you can take it off line, and replace it without taking the system down. In some installations, IBM will automatically be notified of this failure and send out a service person to replace it without even a phone call. Try that with an INTEL box of any kind.
The architecture is also very stable, in that the instruction set hasn't changed much over the years. This allows it to be perfected as it grows. You can still run programs from 25 years ago on today's systems (at least in many cases), without recompilation (though it won't be fully optimized).
2. Scalability. Yeah, it's a single box. But even in its infancy it had a processor complex with the option of containing multiple processors, each of which has been designed for this purpose. You can also many mainframes and form what IBM calls a sysplex. Processing power can also be partitioned into what is known as LPARS, meaning multiple (possibily heterogenous)operating systems can be running simultaneously on the same system. Processing power can be split in a specified proportion, not necessarily equally. They also won't interfere with each other because it is not done at a software level.
All this gives Linux a reliable hardware platform, which in turn makes Linux look more credible.
Forget the drug testing stuff. How about the new
credit checks? As recently as two years ago, I used to be able to find a job as a developer with no credit checks. Now, the present company I work for and many others I've seen require this. Oh sure, they give you a form that asks for your consent. However, if you don't give it they don't give you a job! I wish we could all be so wealthy as to stick it to these extortionists, but we all have to eat. I think this is a small case in point for why full disclosure rules are not enough, but I digress.
As a developer, I don't see the value in sharing my credit information to a company. How about determining your hiring decisions based on solid demonstrated experience, education or certifications instead of vague socialogical extrapolations of what 3 bounced checks might mean. Who knows, maybe if companies hired people not on the basis of credit checks the hired persons might IMPROVE their credit.
Privacy is a big issue, though. While I agree that some amount of drug testing has great merit in many situations, it still is an invasion of privacy which yields some reward to the organization, and society in general. Yet, snooping on someone's financial background serves no purpose to society or the hiring organization. Credit checks just create additional exposure of personal information. Remember the old saying--"Three men can keep a secret as long as two of them are dead." At some point a credit profile is going to be exposed to the wrong people, such as criminals, or fellow employees. Criminals could illegally withdraw from checking accounts. Coworkers might discover a bankruptcy, or that your buried deeper than the Titanic in debt. Unlikely, perhaps, but why would you want to chance it?
Slashdot Mirror
Although I use C++ professionally myself, I'm not ready to say it's easy to debug or understand. The C++ templates, while powerfull, are difficult to debug when the projects get large. It can be hard to determine what they expand to. My debugger, in particular, gets confused with them. VSlick gets confused with the preprocessor tags, too.
Second point: I suspect the choice of language might be a Hobson's choice. Either they were told to use it, or must use it for performance considerations.
On Java: You don't fully win here either, because although it may be more portable, it isn't as stable. I don't mean realiable, I mean that Sun keeps on changing it much more frequently. It's also a very vast platform--just look at how many books at Barnes and Noble there are on the subject. Way too much to read.
Compiler: If C++ is going to be used, your choice of compiler could be as important as the library you use. If someone could suggest a good one for Windows that compiles to native windows with native libraries (not cygwin), that's free, I'd be more than interested.
I too wonder this. Having only used RedHat 7.x,
I can't say what the differences are between
distributions. I thought the end result was the same, once you got past the installations of the OS and applications. Not flamebait, but I'm just curious.
I haven't mastered Swing or AWT. I didn't want to spend time reading 100 pages on ListModels and such just to put up a stinking tree control. Is it any simpler in the SWT?
My gripe is, that I really miss GUI programming in VB, where things were nice and easy and it didn't require reading 500 pages.
Vi. Admirable.
Might I suggest VIM!
What advanced storage techniques? If there are such, they need to be more prominently published. Also, if XML needs to be stored in a more compressed form, why can't XML be passed in a compressed form? Why can't it just be text compiled into a binary encoded format? Use commonly available open source code to read it.
Let's get away from this obsurd notion that everything has to be done in text. It's not faster. It's not more portable. Spelling tags out instead of using symbols doesn't make things platform independent. Just because you can't use a regular text editor doesn't make it harder to use. Just faster.
XML is another attempt by bad software engineering practices to slow down good hardware engineering practices.
Okay, so the marketers have a point. But how about at least having a plain option. Low frames. Low or no images. Or putting a description on images which act as buttons or such.
Remember, the more you embed your information in fancy web stuff, the harder it is to find it in a web search.
The idea that displaying text, excuse me, text with layout or markup tags, being slower than molases is not impressive. Does anyone else cringe when they have to open up a help link web link from a local application?
It takes really long because browser people insist on having a JRE, javascript interpreter, sometimes mail and newsreaders. Of course then there's the plugin's.
Whatever happened to the good ole days where text was text, and help files where man pages or a Windows help file. And desktop publishing was done on WordPerfect?
Yes. Yes. Yes. Do get angry at these web people. I used to be able to dial directly into my bank and download my transactions, and pay bills, all without a web browser. And it was faster. I don't care what you web people say. Life is faster when you don't spell everything out in plain text and use pretty graphics and javascript and such.
Yes. Get rid of the excessive javascript, or even better, don't use it at all! Get rid of the excessive pictures. Don't put a back picture when I could use my back key! Don't create popup menus, just use links. Don't put up ads on bank account pages, especially after the customer has paid you $6.95 per month.
And give the information! Don't make us email you for it. Don't make us call some 800 number and talk to a salesperson. If you have prices, put them up! Don't hide them unless you're ashamed of them.
Have honest links. If you have a download link for an application, for instance, don't make us go through 10,000 slow, image laden web pages just to download the thing. A download link should take us to a downloadable file! (Or a page with the OS selection and such). Forget the mirrors crap. Just ask us a location and direct us to it.
To the web developers: Make life simpler, and faster. Not slow and annoying!
I disagree about Flash. I really wish web developers would have the courtesy of not using things like this. The web protocol and most browsers with them, is really slow. It's also not innovative except in allowing people to pass whole words in the form of tags when they could pass symbols and save bandwidth. We don't need to make it any slower. So if using a standard such as Mosaic 1.0 saves bandwidth by cutting out the fancy crap, I'm all for it. I don't use the web for pretty pictures. I use it for research, and people who insist on developing software for the the absolute slowest GUI available.
What do you mean StarOffice being adequate? It's just cheap.
I find the opposite to be true, except for Postgresql.
Unix=Windows-(Ease of Use+Decent OfficeApplications+Hardware drivers and support+Programmers who aren't cockier than their code deserves+Internet Explorer which face it, is still really good, inspite of you MS haters)
A few points:
Windows isn't going away, and won't be supplanted by a Unix variant, Linux or otherwise, anytime soon.
Very few large companies have just one platform, and most have established applications on multiple platforms. Switching is cost prohibitive.
*nix is not revolutionary, it's just as much a legacy as mainframes.
The idea of everything related to system and software configuration being stored in slow to parse text, in multiple formats is Unix. Having a single registry, stored without wasting space, is revolutionary.
Further, Windows problems can usually be fixed by commonly available utilities, requiring little skill. Oh, I'm sorry, I forgot that I should be a kernel expert to maintain a *nix system.
*nix programs are poorly named, either cute or short, causing much confusion. Task Scheduler vs crond or anacronda. Of course we all know about the thousands of text readers/displayers, etc, such as nroff, xv, latex (I know the case is wrong).
Support and software availability is better on the Windows platform. OK, yeah I can get StarOffice or OpenOffice for Linux. But I can also get that on Windows, plus MS Office which is better than anything on Unix systems, plus WordPerfect (built for Windows and emulated on *nix systems), and Lotus SmartSuite.
All of the above is a different story of course, if you're comparing the two platforms for purely server use. In that case you would probably pick *nix, but not necessarily.
If PC's are fairly cheap, then why not just buy a 2nd PC and put Windows on it. You're going to have to buy it anyway. Might as well have the real thing.
A few comments:
There's two aspects to depreciation, however. From an accounting/tax perspective, accelerating depreciation is preferable, since it reduces taxable income faster. There is a school of thought that says depreciation is almost a fiction from a cash perspective, since the expense isn't a cash expense (at least not until the next purchase).
Netscape 4.x can't display pages clearly because whoever is responsible for X windows makes it a challenge to change the fonts used with it. Styles are window manager issues, not browser issues. Probably most of the interface problems are X problems, which though thoroughly worth griping about, aren't relevant. The last I tested Netscape 6, I think it had the same problems.
Fortunately for me, I've been using IE, which seems to be more stable then the above. It really helps to be using the hedgemonic browser. I'm not sure how these people got their IE browsers to crash, but I certainly remember Netscape crashing perpetually on Linux, since I had to muck with the X and Gnome/KDE settings to get it to work. I never have to do this with Windows, because It Just Works! And faster at that!
The reviewer did point out a few common needs every browser should have though:
Javascript popup disable feature -- Mandatory, mandatory, mandatory.
Image disablement. But let the user choose which ones to disable, such as banner adds.
Speed and simplicity. Stop trying to cram everything into one "browser."
It's nice to use a legible, clean and intelligently laid out web page. But flashy graphics, macromedia flashplayer crap, no thank you. Maybe all web developers should use telephone lines--see how they like all those frames, graphics only buttons, etc. Oh, and stop using that Adobe Acrocrap stuff. Just use HTML like the rest of the page.
Sure you could run it through a compressor, but how often would that happen? I thought people actually like being that verbose and inefficient.
A Few Points, which are almost never mentioned, because they're blasphemous to most web users. But first, I realize ASN.1 is a syntax specifying language used for intellegent information passing, unlike XML/XHTML/HTML which is for bloated, inefficient message passing. However, the, for me, point is speed, not syntax. 1) Its about time someone realized that spelling everything out into words and wordlike symbols actually costs something. Admit it. It takes longer to pass delimiters taking the form of then to come up with a byte encoded form of transmission. 2) Words are not more portable. I can understand a binary bitstream on an intel platform the same as on some Sun machine and even on an IBM mainframe using EBCDIC as long as I knew what it's codepage was. I don't need things spelled out to transfer it to a different hardware or OS platform. Try it sometime if you don't believe me. 3) Wordlike symbols are harder to translate. Its easier to attach multiple words, all meaning the same thing in different languages, to one byte encoded symbol than to an English like word. 4) Parsing speed. This is relevant because some people think XML is a good distributed message passing scheme between two machines--read machines, not human beings. It also doesn't require extra, exotic, hard to find, libraries either. Oh, and stop saying computers are getting faster as though that were a justification. These poor hardware engineers are having a tough time keeping up with incompetent software engineers. With that attitude, the only reason computing might get faster is because of the hardware, certainly not the software. These people who say this need to go back to a 80386 so they can learn to program with resource restrictions--which produces good programs on faster machines. 5) Who cares if a human can read it. It takes special XML readers to format documents in the intelligent way they're meant to be. And those readers are really hard to find if you're cheap like me. Just try to do a search for "Free XML readers". You'll come up with thousands, most of which are just irritating shareware, or slow Java based programs. Further, humans can read all forms of communication, binary or otherwise, just use a debugger if neeeded. Most text editors have a hex viewing/editing capability. Yes, you'll have to do some thinking, and it will be slower to decipher, but it will be much SMALLER. 6) You don't need XML/HTML/XHTML or like markup languages to have a standard. You could standardize on a binary encoding. Yes, maybe we could have a lab experiment on this. I think we really could! Again, try to get people to agree on a syntax derived from human language. It's easier to pick an octet stream and let people decide what it means in their language. I don't thoroughly hate XML/HTML/XHTML, or like ideas. People just misuse and over use them. It
Amen.
Amen. Amen. Amen. Let me begin my rant! This problem isn't limited to Java though. How many times hasn't that frustrated all of us. It's these irritating web jerks who have to make us surf through 80 billion links which all say "download now" just to finally get to the page. Plus, most sites usually want an abundance of information to get the "free" download, and another goddamn password. Then they have to give me a laundry list of download sites because no browser maker has yet to automate the process of finding a reliable site from which to download. Also, we don't all have T1 lines going into our houses. Cable modems in my area cost $40+/month. Five megs for a JRE alone IS huge, and it would be nice for MS to provide this for that reason alone. I'm sorry if I don't enjoy spending 3 hours babysitting a modem connection. This all leads to my conspriracy theory that all software makers try their absolute hardest to slow down computers to make sure that they always run at the speed of a 286. Why else would we have things like Java, XML and the web?
Actually I used to work on a mainframe myself. OS/390 v2.4-2.8. And yes, I have used JCL rather extensively, and have found it do what you say--but not much more. It's a very precise way of executing files, and crude file management. Compare that with shell scripts which don't require column alignments and a // in column 1. Its advantage is simplicity in reading once you understand it, because you can't do much with it.
OS/390 is a really good OS internally. I just don't like everything being limited to units of 8 characters (PDS member names, portions of a DSN,etc). You also can't beat it for programs like RACF for power. With the Unix System Services (which includes vi!), it's gotten a lot more flexible and enables easier porting of UNIX programs. However, I've done admin work on them, from setting up RACF classes, users, TCP/IP stacks, some SMP/E, parmlib setups, etc. Reliable? Definately. Powerful? Usually. Easy? Almost never. Overall though, the architecture and design would make most people here on Slashdot feel in awe of it. Incidently, most of these complaints don't apply to VM, which is what Linux first ran on. I think I saw an article saying IBM got 40,000 partitions running.
I stopped working on them because my product didn't want any more Unix type work done, not because of hatred. I respect them quite a great deal.
I can sympathize with coworkers who pawn off work. Unfortunately, they don't need to be drunk to do this.
If a person was drunk you'd think that would show up in there prior experience, and perhaps the drug test. Sure they can sober up for an interview or a drug test, but how are you going to find that out in a credit check?
Secondly, we already tell private businesses how they must conduct hiring practices. In particular, how they may not conduct it. I'll argue that it's good to have anti-discrimination legislation. I've talked with hiring managers who say it's illegal to ask a company listed as a reference for more than dates of employment, and salary/wages. This is to prevent blacklisting. So if you can't get that information, which would be much more directly relevant to a hiring decision, why should you get the highly indirect/speculative and personal financial background of a candidate?
Further, companies and organizations have a right to keep trade secrets, and such, but why privacy? Companies and organizations are by nature far more exposed to the public than people are. For instance, financial statements are public in the case of corporations, and some companies. In short, private businesses aren't as private as you may think, but people are, and need protection.
I agree that this increases prestige for Linux. Mainframes have gotten an extremely bad reputation for several reasons that Linux would address.
1. Ease of use: Mainframes have historically been hard to use and/or interface with. Anyone who has worked with MVS or what is now known as OS/390 and next Z/OS would remember horrible things. Examples are JCL (Satan's version of a batch file or shell script), Record based file systems such as PDS's, VSAM datasets, and the like which either can't or are hard to directly copy or interface with, and must be preallocated before use. No GUI interface, and limited command line capability in ISPF. There is an X Windows capability via the UNIX services layer, but applications and utilities haven't been ported there much.
2. Culture--This is good and bad. Mainframe programmers tend to be very disciplined. They can also be resistent to change (ie. stubborn), slow because they're disciplined or because they're just slow.
Yet these are just the OS and cultural problems. This is what mainframe architecture gives and has given almost since conception.
1. Reliability. If your DASD(permanent storage), CPU or other device goes out, you can take it off line, and replace it without taking the system down. In some installations, IBM will automatically be notified of this failure and send out a service person to replace it without even a phone call. Try that with an INTEL box of any kind.
The architecture is also very stable, in that the instruction set hasn't changed much over the years. This allows it to be perfected as it grows. You can still run programs from 25 years ago on today's systems (at least in many cases), without recompilation (though it won't be fully optimized).
2. Scalability. Yeah, it's a single box. But even in its infancy it had a processor complex with the option of containing multiple processors, each of which has been designed for this purpose. You can also many mainframes and form what IBM calls a sysplex. Processing power can also be partitioned into what is known as LPARS, meaning multiple (possibily heterogenous)operating systems can be running simultaneously on the same system. Processing power can be split in a specified proportion, not necessarily equally. They also won't interfere with each other because it is not done at a software level.
All this gives Linux a reliable hardware platform, which in turn makes Linux look more credible.
Forget the drug testing stuff. How about the new credit checks? As recently as two years ago, I used to be able to find a job as a developer with no credit checks. Now, the present company I work for and many others I've seen require this. Oh sure, they give you a form that asks for your consent. However, if you don't give it they don't give you a job! I wish we could all be so wealthy as to stick it to these extortionists, but we all have to eat. I think this is a small case in point for why full disclosure rules are not enough, but I digress. As a developer, I don't see the value in sharing my credit information to a company. How about determining your hiring decisions based on solid demonstrated experience, education or certifications instead of vague socialogical extrapolations of what 3 bounced checks might mean. Who knows, maybe if companies hired people not on the basis of credit checks the hired persons might IMPROVE their credit. Privacy is a big issue, though. While I agree that some amount of drug testing has great merit in many situations, it still is an invasion of privacy which yields some reward to the organization, and society in general. Yet, snooping on someone's financial background serves no purpose to society or the hiring organization. Credit checks just create additional exposure of personal information. Remember the old saying--"Three men can keep a secret as long as two of them are dead." At some point a credit profile is going to be exposed to the wrong people, such as criminals, or fellow employees. Criminals could illegally withdraw from checking accounts. Coworkers might discover a bankruptcy, or that your buried deeper than the Titanic in debt. Unlikely, perhaps, but why would you want to chance it?