Tell me, can you sit down when you talk out of your ass?
"Yeah, because Sun's "organizational structures" for open source projects have been such huge successes, right?"
Yep. Ever hear of NFS? NIS? More to the point, take a look at the OpenSolaris community and tell me what's wrong with the organisational structure. It's very similar to the standard open-source structure, except that it addresses some of the problems that have cropped up in that model (fragmentation, dead projects).
I'm not even going to address your fairy-dust predictions about licenses. Entertaining though.
"Linux already has tracing technologies and it has multiple excellent file systems"
Linux built-in tracing is a poor cousin to Unix truss. Such specialised projects as LTT provide significant advancement, but nothing even close to what Dtrace provides. Complete system probes are something completely new in the world of computing, with the possible exception of some _very_ esoteric realtime systems. This is actually big news, and Linux should be absorbing it, rather than sneering. Ditto for ZFS. Even with some of its features not yet implemented in release code, it is the first significant filesystem mindset shift to happen in decades. Linux' "multiple excellent file systems" are nothing special--the best of them work acceptably, but that's it. ZFS will transform small-storage computing in a few years.
OpenSolaris isn't going to be a dud because Solaris10 is already a hit, and the serious opensource community (i.e., the programmers and not the whiners) are already working on OpenSolaris--often in parallel with other OSes.
OK, I'm a hardcore, fulltime, Solaris admin. I've been heavily involved in rolling out Solaris 10 to the production systems at a major oil corporation. In other words, I'm a bit biased.:-)
Be that as it may, let me make a few comments.
1) Nevada is development code, not release. If there aren't bugs in your dev. code, then you're either the finest programmer since about 1960, or you're not doing anything. 2) OpenSolaris in general is not the place to go for release code; It's the community work, warts and all. If you want a production OS, you use Solaris10. 3) Having said that, let me also add that Solaris10 is documented. Heavily. Coherently. Completely. HPUX and AIX are close, Linux isn't even an also-ran in the documentation realm.
So let me talk about some of the good and bad we're seeing with Solaris10 in the real world.
Let me start by stating that dtrace rocks. Most admins don't write scripts in it, as you suspect--however, they do download them from programmers who give back to the community. Similarly, zones rock too--companies are using them to compartmentalise their environments (for example: one database instance per zone), which makes migration between machines a trivial process. BrandZ is an interesting offshoot, but is likely to be less important for users than for developers.
Hardware support (specifically non-Sun, x86/x64) hardware support is amazing. Really, Solaris will work on anything!!!
And if you believe that, I have a bridge to sell you.
OK, let's come clean. Hardware support is rapidly um... sucking less. It's still nowhere close to the Linux ballpark, and probably won't be for at least three years. That's not surprising from a company that continually tried to kill off their X86 offering for several years, before rather suddenly committing to it. Let me come back to this in a moment.
I'll readily admit that the anti-Linux sentiment is very strong in the (Open)Solaris world, but you should understand where some of the frustration comes from. Daily (hourly!), the various newsgroups and discussion forums receive posts that come across as, "I can do THIS in Linux, but I can't in Solaris. Why does Solaris suck so badly?!" The answer is usually that Linux has some nonstandard (and more than occasionally undocumented) extensions to standard Unix tools. In other words, this 'thing' will not work in Irix, AIX, HP-UX, *BSD, OSF/1, OS X, or any other Unix variant--only in Linux. Furthermore, if that behaviour is really necessary (it rarely is), then the tool is probably available as a source or binary download to anyone interested.
I can't comment on ACPI, other than to state that I have never used a computer for any length of time, running any OS, that did power management properly. That includes Linux (RHEL3 and older), Windows (XP and earlier), or Solaris (10, etc.)
Don't get me wrong here--Solaris on commodity hardware still has a ways to go. However, Solaris on Sun hardware (either SPARC or X64) is the best thing going in computing right now. For those two reasons, OpenSolaris really does have the potential to take the world by storm next year. The community has been presented with both a challenge (make this a true commodity-hardware OS), and a clear goal (behaving like Solaris10 on Sun gear). Furthermore, since Sun is feeding contributions back from OpenSolaris into Solaris, the 'official' OS will continue to get better.
In other words, the OpenSolaris community will thrive because there's an intriguing challenge facing them, and a clear reward as a result.
Well done--that was one of the best posts I've read in a long time. Naturally I'm going to have to argue with it.:-)
The comment about "# _" being the most exciting thing to see for a hacker (in Unix at least) is right--and that's not only boring, but arcane. Explaining the arcane is one of the prime purposes of sidekicks in movies. Imagine:
"Hey, check it out--I'm in!" "What do you mean, you're in? It looks like the same crap you've been staring at for six straight hours." "No, look--that prompt means I'm root--I can do anything at all on this machine! Don't you see?" (pause)
OK, so I'm no screenwriter. Still, having a sidekick around is a great excuse for explaining stuff to the audience. Furthermore, it's possible to gloss over the details (sins of omission) rather than invent a big pile of crap for the technology aspects of a movie.
Every item on that list was geeky--most just didn't happen to be geeky in a computer-centric fashion.
A few points here: 1) Pod espresso machines are NEVER as good as you can make with really fresh, properly ground coffee, in a good machine, with some significant practice. At the same time, they're probably not as bad as most people are likely to get, since people are lazy. On the other hand, espresso geekery is absolutely first-rate. 2) Knife hardness != knife sharpness. 3) Subwoofers should be purchased with the speakers, and should be designed together. 4) Actually, I'm going to stop. This was just a crappy article all 'round.
What do you mean "now?" The RIAA has worked contrary to the interests of artists for decades--possibly since their inception. They're a parasitic organisation that steals money from creative people.
I haven't gotten over Gracenote's theft yet. I typed in about 100 albums worth of data (out of my 400 album collection) and submitted it to the FREE CDDB. Gracenote stole my work and resold it.
But what's there to say here, really? Steve is a lying, two-faced asshole who probably actually feels wronged by all of the bad press he's earned. He belongs in the same category as Jeff Immelt and Sanford Wallace. In the same jail cell as well, actually.
In junior high I was always trying to come up with something cool for the science fair. In grade seven, it was a polarimeter for monitoring stress. In grade eight, it was a cloud chamber. In grade nine, I built a fully functional van de graff generator, only 15cm high. I had a plan for "The smallest van de graff generator in the world!" but made a mistake in measuring, and actually made mine slightly smaller.
That thing was fantastic. Still in my parents basement, somewhere.
So my first thought is this: If Bush's behaviour is treason (and as an outside observer, I tend to agree), then what are you personally doing about it? Have you phoned your federal representative? Have you tried to get a motion of impeachment started?
In short, it ain't treason if the populace doesn't hold you accountable.
Secondly, Bill has done everything that Bush has, except acted out of fear. He has lied, cheated, swindled, and stolen. His realm is money, Bush's is world power. Not a lot of difference.
We routinely get security warnings from our security team about horrible flaws on our Unix systems that Symmantec has flagged. Usually they're along the lines of,
"You're running Solaris 8. The default install of the first release of Solaris 8 had an obscure FTP bug and so YOU'RE GOING TO GET HACKED!!!!" This bulletin applies to every release of Solaris 2.5.1, 2.6, 2.7, 7, 8, 8.1, 9.0, 10; HP-UX... Linux...AIX...NCFTPD...WU-FTPD..." and so forth.
Usually it's a single bug in a single version that we've already patched, if it's a service that we actually use (given that we run a pretty tight network internally already).
Luckily our security team is smarter than the software they use, and understand the explanations we give them. Once in a (long) while, they actually find something that needs patching.
Well, I guess we're going to have to agree to disagree--at least I will. I understand your points and mostly agree with them, but disagree with the conclusions. I'll just address your guess at the end, and maybe you'll understand a bit more about where I'm coming from.
"I'm going to guess that you're a young, idealistic kid at university. You have very little experience with the industry or its history, and you don't realize that all of your "big ideas" and predictions have been thought of and predicted already. A lot of bright people have thought about these topics, from Alan Turing and Herbert Simon, to more practical folks like Sergey Brin and Bill Gates. Before you sound off, how about reading what they have to say?"
Exactly the opposite, actually. University was just a smidge under 20 years ago for me, and although my degree was in an entirely different field, I've been in computing long enough to see exactly what I'm talking about. When I started working with computer geeks, a sysadmin had to be a programmer for several years first. Later on, a good admin could program but it wasn't necessary beyond solid scripting skills. That's now dwindled to the point that being able to troubleshoot a failed autoconf is a rarity. No problem, except that the skills which are replacing the 'obsolete' ones are less creative and more along the lines of 'vendor technician.' Gaming is becoming less creative at an insanely fast rate, as technical barriers are removed. Why invent something new when you can just make a game more like reality than it was last year?
I think we're on the same page. I'm just more pessimistic than you (and many others).
The only thing worse than an aggressively ignorant idiot is a foul-mouthed one.
In the 1830s, it was discovered that some materials acted as neither pure conductors nor pure insulators. They called them semiconductors, and they were a curiosity until the 1890s, when they were found to be useful as rectifiers and photovoltaic cells. Another 40 years later, and people started to consider them as a replacement for the triode vacuum tube, which was immensely useful but fragile and difficult to deal with.
Pure research in new directions isn't just allowed because it 'might lead to something,' it's absolutely essential in order to progress beyond refinement of the existing.
First of all, it's called RESEARCH! It's a very new and different bit of science--who knows where it could lead us?
Secondly, just because things are a pain in the ass doesn't mean they don't have useful applications. NMR/MRI have been dependent on low-temp superconductors (i.e. liquid He or even colder) for decades, and they're immensely important for research and medicine.
Re:What about the Microsoft "head-fake?"
on
In Search of Stupidity
·
· Score: 2, Insightful
I think you need to fast-forward a little bit, to somewhere about 1994. OS/2 2.0 was out and recognised in the computing industry as superior to Window NT 4.0. NT5 was still a long ways away, Win3.1 was still a GUI on top of DOS, and the Pentium floating point bug was just getting noticed. That's when IBM announced OS/2 2.1--a really great next generation OS, for the home and also for the office. It had flash, it had utility, it actually ran Windows programs, and IBM was going to start selling PCs with the brand-new PPC chip, SCSI hard drives (like Apple), running OS/2. Microsoft and Intel were faltering badly, and IBM had a perfect coup in the works.
I ended up talking quite extensively with an IBM architect at the time. I was impatiently biting my nails, waiting to spend all of my money on a fancy new legacy-free DOS-free IBM computer which would kick the ass of anything else out there. The comment I got back from IBM was, "Well we've got a few things to polish still, so we won't release anything until we're ready. The market will wait for us, if the product is good."
I couldn't believe I was hearing this. The market NEVER waits for the product--least of all in computing. If there's real competition, the first to market wins. Not the best, not the fastest, not even the prettiest, but the best marketed--and the first step in marketing is get to the customer before your competition.
IBM blew it. They completely failed to capitalise on a chance to make a huge change in the PC marketplace. It wasn't Microsoft, it was IBM.
Take a look at another reply for an example of how some of my thoughts will be flushed out.
Now, your facetious suggestions of cron and compilers are in fact
exactly
what I'm talking about. Programming started out with tube arrays, and went to toggle switches and drums and eventually assembler, which was essentially low-level compiling. When symbolic compilers came along, they made coding VASTLY easier--and we created more complex code as a result. Higher level compilers led to more complex code, metacompilers, and so forth. We are, right now, at the point where the code for major programs can't be understood by any single individual (just like microprocessor design, in fact). We are very close to the point where no group or team will be able to understand the entire code for a major program--the result is that we'll create compilers that create code to be fed into compilers, which create lower-level code, which will be fed into compilers which create machine code. Eventually, the "programmer's code" will be no more than a concise description of what the program should do. (This is a good thing, by the way.) At that point, programming as a field of progressive science will dwindle, and become a technical trade, like riveting or welding or house framing.
As far as self-administering computers, cron
was
exactly that, compared to having to manually schedule jobs or program delays. In fact, it could be considered to be the first example of self-administration. Nowadays however, we have computers that detect progressing faults (memory, CPU, etc.), offline the questionable hardware, and place a service call to get a part delivered. We also have similar items for software--Consider some of Sun's recent technology: ZFS self-administers much of the task of disk management, and services provide intelligent self-healing software (i.e. the OS will detect the failure of the application, clear faults if present, try to restart it, and then fail only if it doesn't start properly). HSM is suffering a rennaissance too, under the guise of "information lifecycle management."
Look at it all another way: The technology we have RIGHT NOW, with no further advances, will allow us to create games not with a masters in computing science, nor with a bachelors degree or a technical diploma, but rather with a background (degreed or not) in arts and human studies.
Or look at it in yet another way: A master's of engineering in games and media integration? This isn't engineering, and shouldn't be. It's an art-heavy interdisciplinary field, and will shift farther away from engineering as time goes on.
Open Source shouldn't be capitalised. It shouldn't be used to categorise software. It shouldn't be a philosophy, an ideology, a theology, an attitude, or a way of life. Nor should it be a deciding factor in chosing software.
What open source software SHOULD be is one set of conditions that applies to software you may want to use for other reasons. Decide what to do for technical and business reasons, not ideological ones. If the tool best fits your needs, then figure out if the support and business models fit your needs as well.
As soon as you start talking about Open Source as a separate entity, you're evangelising. Consider this comment:
"What we are looking to do is: formalize and consolidate our global Open Source knowledge to accommodate new and existing client requirements; define a review process that will enable us to quickly review Open Source tools, applications, and so forth"
Now get rid of Open Source in there:
"What we are looking to do is: formalize and consolidate our global knowledge to accommodate new and existing client requirements; define a review process that will enable us to quickly review tools, applications, and so forth"
It's just as clear, just as concise, but doesn't pretend that "Open Source" is something magic and different from ALL OTHER TOOLS out there!
Get over open source as a special category, and just use the tools you need. Contribute if you can. Quit believing that it's special.
Here's a controversial suggestion: computing is a dead-end field.
Not today, not tomorrow, and maybe not for a decade or even two, but I honestly think that computing as a career will drop to the same level as janitor. First, computers will become more-or-less self-administering, eliminating one sector. Later, code will become too complex to write for people (already people can only write small subsets of a program), so computers will take over generating code. Exuent programmers. Creative expression requires unknowns to be useful. As ideas are created and discarded, we will get to a standard way of making things, which will leave computer creativity to the font of all other branches--with the artist.
Computing as a creative technical field is going to die, and probably by the middle of your career, if not sooner.
Seriously, you've got a point. An honest OS vendor guarantees backwards compatability only using published, official programming interfaces. Application vendors turn around and find some obscure undocumented library call that makes life easier for them, and use it everywhere. When the OS is upgraded, the app breaks and "backwards compatability" gets shat upon unfairly.
A major problem though, is that many of these applications are niche products which makes them a de facto monopoly. If the customer needs functionality "x", then they're the only game in town. Thus the customer says, "I don't care if it's the application vendor's fault, fix it in the OS or I'm switching to Windows!"
Result: Backwards compatability is guaranteed for cases where it _should_ break. Unpublished interfaces become official and supported standards. Cruft carries forwards.
As for the "don't upgrade your computers," that's lovely--as long as you don't want any support for your OS. Kind of hard to get (unrelated) patches for OSes that have been EOSL for four years, though. Next year, a significant part of the western world is going to change daylight savings time, which will require a patch. Any machines that are running old OSes are going to be unable to keep proper time. Pretty close to a metaphorical gun, I'd say.
This is just life, going between OSes. I spend 95+% of my time on Solaris. In the last two weeks, I've had to work on OSF/1, HPUX 11i, and HPUX 9.04. You get used to switching between variants, and start to get a feel for where things are. ("Oh, that's in/sbin instead of/etc, so that means that THIS should be in...")
Trust me, it's good for the soul. And ZFS rocks. (Not to mention zones, dtrace, and now brandz).
Forget about the fly-by-night CEOs. Consider taking out the long-term criminal ones. Jeffrey Immelt of GE is a filthy hypocritical liar, and one of his cronies, John Bucci, is rising higher and higher in the company by lying to his employees and customers, in order to cheat them.
Then there's Sony, trying to screw customers in order to pump up the short-term stock options of the execs. These are all people who should be placed carefully in sealed plastic bags.
It's not just one company or a few, though. Once a company gets big enough, this sort of thing is inevitable.
Slashdot Mirror
Culture in America was officially declared dead on December 8, 1980. The actual time of death is unknown.
Tell me, can you sit down when you talk out of your ass?
"Yeah, because Sun's "organizational structures" for open source projects have been such huge successes, right?"
Yep. Ever hear of NFS? NIS?
More to the point, take a look at the OpenSolaris community and tell me what's wrong with the organisational structure. It's very similar to the standard open-source structure, except that it addresses some of the problems that have cropped up in that model (fragmentation, dead projects).
I'm not even going to address your fairy-dust predictions about licenses. Entertaining though.
"Linux already has tracing technologies and it has multiple excellent file systems"
Linux built-in tracing is a poor cousin to Unix truss. Such specialised projects as LTT provide significant advancement, but nothing even close to what Dtrace provides. Complete system probes are something completely new in the world of computing, with the possible exception of some _very_ esoteric realtime systems. This is actually big news, and Linux should be absorbing it, rather than sneering. Ditto for ZFS. Even with some of its features not yet implemented in release code, it is the first significant filesystem mindset shift to happen in decades. Linux' "multiple excellent file systems" are nothing special--the best of them work acceptably, but that's it. ZFS will transform small-storage computing in a few years.
OpenSolaris isn't going to be a dud because Solaris10 is already a hit, and the serious opensource community (i.e., the programmers and not the whiners) are already working on OpenSolaris--often in parallel with other OSes.
OK, I'm a hardcore, fulltime, Solaris admin. I've been heavily involved in rolling out Solaris 10 to the production systems at a major oil corporation. In other words, I'm a bit biased. :-)
Be that as it may, let me make a few comments.
1) Nevada is development code, not release. If there aren't bugs in your dev. code, then you're either the finest programmer since about 1960, or you're not doing anything.
2) OpenSolaris in general is not the place to go for release code; It's the community work, warts and all. If you want a production OS, you use Solaris10.
3) Having said that, let me also add that Solaris10 is documented. Heavily. Coherently. Completely. HPUX and AIX are close, Linux isn't even an also-ran in the documentation realm.
So let me talk about some of the good and bad we're seeing with Solaris10 in the real world.
Let me start by stating that dtrace rocks. Most admins don't write scripts in it, as you suspect--however, they do download them from programmers who give back to the community. Similarly, zones rock too--companies are using them to compartmentalise their environments (for example: one database instance per zone), which makes migration between machines a trivial process. BrandZ is an interesting offshoot, but is likely to be less important for users than for developers.
Hardware support (specifically non-Sun, x86/x64) hardware support is amazing. Really, Solaris will work on anything!!!
And if you believe that, I have a bridge to sell you.
OK, let's come clean. Hardware support is rapidly um... sucking less. It's still nowhere close to the Linux ballpark, and probably won't be for at least three years. That's not surprising from a company that continually tried to kill off their X86 offering for several years, before rather suddenly committing to it. Let me come back to this in a moment.
I'll readily admit that the anti-Linux sentiment is very strong in the (Open)Solaris world, but you should understand where some of the frustration comes from. Daily (hourly!), the various newsgroups and discussion forums receive posts that come across as, "I can do THIS in Linux, but I can't in Solaris. Why does Solaris suck so badly?!" The answer is usually that Linux has some nonstandard (and more than occasionally undocumented) extensions to standard Unix tools. In other words, this 'thing' will not work in Irix, AIX, HP-UX, *BSD, OSF/1, OS X, or any other Unix variant--only in Linux. Furthermore, if that behaviour is really necessary (it rarely is), then the tool is probably available as a source or binary download to anyone interested.
I can't comment on ACPI, other than to state that I have never used a computer for any length of time, running any OS, that did power management properly. That includes Linux (RHEL3 and older), Windows (XP and earlier), or Solaris (10, etc.)
Don't get me wrong here--Solaris on commodity hardware still has a ways to go. However, Solaris on Sun hardware (either SPARC or X64) is the best thing going in computing right now. For those two reasons, OpenSolaris really does have the potential to take the world by storm next year. The community has been presented with both a challenge (make this a true commodity-hardware OS), and a clear goal (behaving like Solaris10 on Sun gear). Furthermore, since Sun is feeding contributions back from OpenSolaris into Solaris, the 'official' OS will continue to get better.
In other words, the OpenSolaris community will thrive because there's an intriguing challenge facing them, and a clear reward as a result.
Well done--that was one of the best posts I've read in a long time. Naturally I'm going to have to argue with it. :-)
The comment about "# _" being the most exciting thing to see for a hacker (in Unix at least) is right--and that's not only boring, but arcane. Explaining the arcane is one of the prime purposes of sidekicks in movies. Imagine:
"Hey, check it out--I'm in!"
"What do you mean, you're in? It looks like the same crap you've been staring at for six straight hours."
"No, look--that prompt means I'm root--I can do anything at all on this machine! Don't you see?"
(pause)
OK, so I'm no screenwriter. Still, having a sidekick around is a great excuse for explaining stuff to the audience. Furthermore, it's possible to gloss over the details (sins of omission) rather than invent a big pile of crap for the technology aspects of a movie.
Every item on that list was geeky--most just didn't happen to be geeky in a computer-centric fashion.
A few points here:
1) Pod espresso machines are NEVER as good as you can make with really fresh, properly ground coffee, in a good machine, with some significant practice. At the same time, they're probably not as bad as most people are likely to get, since people are lazy. On the other hand, espresso geekery is absolutely first-rate.
2) Knife hardness != knife sharpness.
3) Subwoofers should be purchased with the speakers, and should be designed together.
4) Actually, I'm going to stop. This was just a crappy article all 'round.
What do you mean "now?" The RIAA has worked contrary to the interests of artists for decades--possibly since their inception. They're a parasitic organisation that steals money from creative people.
I haven't gotten over Gracenote's theft yet. I typed in about 100 albums worth of data (out of my 400 album collection) and submitted it to the FREE CDDB. Gracenote stole my work and resold it.
But what's there to say here, really? Steve is a lying, two-faced asshole who probably actually feels wronged by all of the bad press he's earned. He belongs in the same category as Jeff Immelt and Sanford Wallace. In the same jail cell as well, actually.
...and in junior high, no less.
In junior high I was always trying to come up with something cool for the science fair. In grade seven, it was a polarimeter for monitoring stress. In grade eight, it was a cloud chamber. In grade nine, I built a fully functional van de graff generator, only 15cm high. I had a plan for "The smallest van de graff generator in the world!" but made a mistake in measuring, and actually made mine slightly smaller.
That thing was fantastic. Still in my parents basement, somewhere.
Oh yeah. Then he could inherit our egalitarian-but-still-broken healthcare system.
If the US moves up here under Bill or anyone else, you'll suddenly see a whole bunch of passivists taking up arms.
So my first thought is this: If Bush's behaviour is treason (and as an outside observer, I tend to agree), then what are you personally doing about it? Have you phoned your federal representative? Have you tried to get a motion of impeachment started?
In short, it ain't treason if the populace doesn't hold you accountable.
Secondly, Bill has done everything that Bush has, except acted out of fear. He has lied, cheated, swindled, and stolen. His realm is money, Bush's is world power. Not a lot of difference.
We routinely get security warnings from our security team about horrible flaws on our Unix systems that Symmantec has flagged. Usually they're along the lines of,
"You're running Solaris 8. The default install of the first release of Solaris 8 had an obscure FTP bug and so YOU'RE GOING TO GET HACKED!!!!" This bulletin applies to every release of Solaris 2.5.1, 2.6, 2.7, 7, 8, 8.1, 9.0, 10; HP-UX... Linux...AIX...NCFTPD...WU-FTPD..." and so forth.
Usually it's a single bug in a single version that we've already patched, if it's a service that we actually use (given that we run a pretty tight network internally already).
Luckily our security team is smarter than the software they use, and understand the explanations we give them. Once in a (long) while, they actually find something that needs patching.
Foreign student breaks law, gets deported.
That's it. The rest is minutae.
This is how I interpreted the article.
"So what if we infringe on patents? We're Open Source--you can't hurt us!"
Lovely excuse for breaking the law. It certainly should convince more companies that they need to start using this illegal-but-untouchable code.
OK, this is a bit out there, but bear with me.
Couldn't they maybe...make an original movie?
Well, I guess we're going to have to agree to disagree--at least I will. I understand your points and mostly agree with them, but disagree with the conclusions. I'll just address your guess at the end, and maybe you'll understand a bit more about where I'm coming from.
"I'm going to guess that you're a young, idealistic kid at university. You have very little experience with the industry or its history, and you don't realize that all of your "big ideas" and predictions have been thought of and predicted already. A lot of bright people have thought about these topics, from Alan Turing and Herbert Simon, to more practical folks like Sergey Brin and Bill Gates. Before you sound off, how about reading what they have to say?"
Exactly the opposite, actually. University was just a smidge under 20 years ago for me, and although my degree was in an entirely different field, I've been in computing long enough to see exactly what I'm talking about. When I started working with computer geeks, a sysadmin had to be a programmer for several years first. Later on, a good admin could program but it wasn't necessary beyond solid scripting skills. That's now dwindled to the point that being able to troubleshoot a failed autoconf is a rarity. No problem, except that the skills which are replacing the 'obsolete' ones are less creative and more along the lines of 'vendor technician.' Gaming is becoming less creative at an insanely fast rate, as technical barriers are removed. Why invent something new when you can just make a game more like reality than it was last year?
I think we're on the same page. I'm just more pessimistic than you (and many others).
The only thing worse than an aggressively ignorant idiot is a foul-mouthed one.
In the 1830s, it was discovered that some materials acted as neither pure conductors nor pure insulators. They called them semiconductors, and they were a curiosity until the 1890s, when they were found to be useful as rectifiers and photovoltaic cells. Another 40 years later, and people started to consider them as a replacement for the triode vacuum tube, which was immensely useful but fragile and difficult to deal with.
Pure research in new directions isn't just allowed because it 'might lead to something,' it's absolutely essential in order to progress beyond refinement of the existing.
First of all, it's called RESEARCH! It's a very new and different bit of science--who knows where it could lead us?
Secondly, just because things are a pain in the ass doesn't mean they don't have useful applications. NMR/MRI have been dependent on low-temp superconductors (i.e. liquid He or even colder) for decades, and they're immensely important for research and medicine.
I think you need to fast-forward a little bit, to somewhere about 1994. OS/2 2.0 was out and recognised in the computing industry as superior to Window NT 4.0. NT5 was still a long ways away, Win3.1 was still a GUI on top of DOS, and the Pentium floating point bug was just getting noticed. That's when IBM announced OS/2 2.1--a really great next generation OS, for the home and also for the office. It had flash, it had utility, it actually ran Windows programs, and IBM was going to start selling PCs with the brand-new PPC chip, SCSI hard drives (like Apple), running OS/2. Microsoft and Intel were faltering badly, and IBM had a perfect coup in the works.
I ended up talking quite extensively with an IBM architect at the time. I was impatiently biting my nails, waiting to spend all of my money on a fancy new legacy-free DOS-free IBM computer which would kick the ass of anything else out there. The comment I got back from IBM was, "Well we've got a few things to polish still, so we won't release anything until we're ready. The market will wait for us, if the product is good."
I couldn't believe I was hearing this. The market NEVER waits for the product--least of all in computing. If there's real competition, the first to market wins. Not the best, not the fastest, not even the prettiest, but the best marketed--and the first step in marketing is get to the customer before your competition.
IBM blew it. They completely failed to capitalise on a chance to make a huge change in the PC marketplace. It wasn't Microsoft, it was IBM.
Take a look at another reply for an example of how some of my thoughts will be flushed out.
Now, your facetious suggestions of cron and compilers are in fact
- exactly
what I'm talking about. Programming started out with tube arrays, and went to toggle switches and drums and eventually assembler, which was essentially low-level compiling. When symbolic compilers came along, they made coding VASTLY easier--and we created more complex code as a result. Higher level compilers led to more complex code, metacompilers, and so forth. We are, right now, at the point where the code for major programs can't be understood by any single individual (just like microprocessor design, in fact). We are very close to the point where no group or team will be able to understand the entire code for a major program--the result is that we'll create compilers that create code to be fed into compilers, which create lower-level code, which will be fed into compilers which create machine code. Eventually, the "programmer's code" will be no more than a concise description of what the program should do. (This is a good thing, by the way.) At that point, programming as a field of progressive science will dwindle, and become a technical trade, like riveting or welding or house framing.As far as self-administering computers, cron
- was
exactly that, compared to having to manually schedule jobs or program delays. In fact, it could be considered to be the first example of self-administration. Nowadays however, we have computers that detect progressing faults (memory, CPU, etc.), offline the questionable hardware, and place a service call to get a part delivered. We also have similar items for software--Consider some of Sun's recent technology: ZFS self-administers much of the task of disk management, and services provide intelligent self-healing software (i.e. the OS will detect the failure of the application, clear faults if present, try to restart it, and then fail only if it doesn't start properly). HSM is suffering a rennaissance too, under the guise of "information lifecycle management."Look at it all another way: The technology we have RIGHT NOW, with no further advances, will allow us to create games not with a masters in computing science, nor with a bachelors degree or a technical diploma, but rather with a background (degreed or not) in arts and human studies.
Or look at it in yet another way: A master's of engineering in games and media integration? This isn't engineering, and shouldn't be. It's an art-heavy interdisciplinary field, and will shift farther away from engineering as time goes on.
It doesn't take a futurist to see that much.
Open Source shouldn't be capitalised. It shouldn't be used to categorise software. It shouldn't be a philosophy, an ideology, a theology, an attitude, or a way of life. Nor should it be a deciding factor in chosing software.
What open source software SHOULD be is one set of conditions that applies to software you may want to use for other reasons. Decide what to do for technical and business reasons, not ideological ones. If the tool best fits your needs, then figure out if the support and business models fit your needs as well.
As soon as you start talking about Open Source as a separate entity, you're evangelising. Consider this comment:
"What we are looking to do is: formalize and consolidate our global Open Source knowledge to accommodate new and existing client requirements; define a review process that will enable us to quickly review Open Source tools, applications, and so forth"
Now get rid of Open Source in there:
"What we are looking to do is: formalize and consolidate our global knowledge to accommodate new and existing client requirements; define a review process that will enable us to quickly review tools, applications, and so forth"
It's just as clear, just as concise, but doesn't pretend that "Open Source" is something magic and different from ALL OTHER TOOLS out there!
Get over open source as a special category, and just use the tools you need. Contribute if you can. Quit believing that it's special.
Here's a controversial suggestion: computing is a dead-end field.
Not today, not tomorrow, and maybe not for a decade or even two, but I honestly think that computing as a career will drop to the same level as janitor.
First, computers will become more-or-less self-administering, eliminating one sector.
Later, code will become too complex to write for people (already people can only write small subsets of a program), so computers will take over generating code. Exuent programmers.
Creative expression requires unknowns to be useful. As ideas are created and discarded, we will get to a standard way of making things, which will leave computer creativity to the font of all other branches--with the artist.
Computing as a creative technical field is going to die, and probably by the middle of your career, if not sooner.
I sense a lot of anger coming from this one.
Seriously, you've got a point. An honest OS vendor guarantees backwards compatability only using published, official programming interfaces. Application vendors turn around and find some obscure undocumented library call that makes life easier for them, and use it everywhere. When the OS is upgraded, the app breaks and "backwards compatability" gets shat upon unfairly.
A major problem though, is that many of these applications are niche products which makes them a de facto monopoly. If the customer needs functionality "x", then they're the only game in town. Thus the customer says, "I don't care if it's the application vendor's fault, fix it in the OS or I'm switching to Windows!"
Result: Backwards compatability is guaranteed for cases where it _should_ break. Unpublished interfaces become official and supported standards. Cruft carries forwards.
As for the "don't upgrade your computers," that's lovely--as long as you don't want any support for your OS. Kind of hard to get (unrelated) patches for OSes that have been EOSL for four years, though. Next year, a significant part of the western world is going to change daylight savings time, which will require a patch. Any machines that are running old OSes are going to be unable to keep proper time.
Pretty close to a metaphorical gun, I'd say.
"Nothing is where I expect it to be."
/sbin instead of /etc, so that means that THIS should be in...")
This is just life, going between OSes. I spend 95+% of my time on Solaris. In the last two weeks, I've had to work on OSF/1, HPUX 11i, and HPUX 9.04. You get used to switching between variants, and start to get a feel for where things are. ("Oh, that's in
Trust me, it's good for the soul. And ZFS rocks. (Not to mention zones, dtrace, and now brandz).
Straight through the brainpan.
It takes a bit of time to set up and will probably be three years before the momentum is there, but it's the only _real_ solution.
Forget about the fly-by-night CEOs. Consider taking out the long-term criminal ones. Jeffrey Immelt of GE is a filthy hypocritical liar, and one of his cronies, John Bucci, is rising higher and higher in the company by lying to his employees and customers, in order to cheat them.
Then there's Sony, trying to screw customers in order to pump up the short-term stock options of the execs. These are all people who should be placed carefully in sealed plastic bags.
It's not just one company or a few, though. Once a company gets big enough, this sort of thing is inevitable.