Point 1 and 2 could be solved by using by using Novell Linux Desktop. Yes, it's boring, but you can't have it both way.
BTW, the menu in my 9.3 installation have a single volume control applet, and it's pretty damn simple to understand. Actually, it's pretty much a copy of the Windows mixer. WTF is wrong with your install ?
Just a thought here... At least, lets hope Zack is smart enough to be following what Microsoft is doing, so maybe he will be on top of things and can get you guys into the next generation of computing, since most everyone hear does the cover ears, eyes and sing la la la when they hear Microsoft is actually doing something good that is going to change the computing world.
You've read it on Slashdot first, folks : hardware accelerated XML presentation layer will change the computing world ! Forget quantum computing, you really more eye candies in the end.
Does such a thing exist ? Is there, somewhere, a software that I can buy that will satisfy all my needs in IT ?
I.T. departments everywhere will keep wondering when FOSS is going to do something to alter the status quo in a meaningful (i.e. outside the server room) way.
Because, of course, the server room is not meaningful to "enterprise software".
FLOSS is, all thing being equal, cheaper and more flexible. So it's not a "zero-sum equation". The specific of your situation might call for a proprietary solution; I won't hold grudge to anybody in such a situation.
I don't understand why you are mixing enterprise and end-users software in the discussion, unless you confuse a word-processor for an "enterprise software". It's two different market entirely, and FLOSS serve the former much better than it does the latter.
You do a lot of broad generalization (FLOSS devs are elitist, FLOSS are crap, etc) but do not get much into the specifics. Suitability to task is a case-by-case issue. I do not think anybody in his right mind ever suggested you use vi as a word processor.
Well documented
Not take a programmer to install...no hidden options
Lots of closed-source commercial fail these requirements too.
Enterprise-level application require professionnal to install and support them. That's a given. Why would it be any different with FLOSS ?
Developer must be open to suggestions (from regular users as well as developers/geeks)
This is impossible with commercial software, except if the supplier is very small or you are a *very* large customer. Otherwise, you never get to make your suggestions to the developpers. You are always shielded from the production by the marketing and sales departement. You might tell your sales rep your suggestions, but what he does with them is impossible to tell. At least, with FLOSS, you have some measure of access to the devs. Wheter or not they will listen to your suggestion is not a given, but if you bring something worthwhile to the table, they might.
Of course, you have to work within their workflow (ie submit your request through the bug tracking system). At least, you know where your suggestions are going.
Possible support contract available (we'll accept SuSE or IBM support if it's available).
Why IBM or SuSE ? Why not hiring a local service firm ? Or developping in-house expertise ?
The nature of FOSS is not to provide "enterprise-level services"; it is to provide software. Now, if you want to buy "enterprise-level services" for FOSS, there is plenty of people willing to sell that you.
Again, what is the problem with FOSS ? That it does not come bundled with "enterprise-level services" for free ? Cry me a river...
Both RedHat and Novell would come to you with marketable products. Remember, if you want shrink-wrap solution, you need to shop for shrink-wrap solution, and Sourceforge do not sell shrink-wrap solution (where RedHat and Novell do). But don't cringe about the price tag, because all that wrapping is costing mhtem oney, and this cost is being passed to you, the customer. Not that there is anything wrong with it, if you already pay for closed-source commercial software.
In the meantime, I make a living installing and supporting the applications you snob, and it cost my clients a fraction of the price of shrink-wrap software. Go figure.
Well, it's useful to computer scientists. It's peripherically useful to computer programmers, but these could be taught a more pertinent curriculum. IMHO. A curriculum that could include stuff like project management, interface design, software testing and, of course, more software engineering.
While I disagree with your point of view regarding the programmer's expected skillset, I must salute your position on drug screen test. Up here in Canada, (almost) nobody ask for screen test. It never happenned to me, but I think the practice is revolting. People must stand up. Otherwise, what next ? DNA sample ?
I agree 100% with everything you just wrote. My problem is not with the challenge level or the usefulness of CS, it is with the expectative that employers (and, to a certain degree, students) have with regard to CS curriculum. If we expect CS graduate to work as programmers, as it is most certainly the case right now, are we preparing them adequately ? The answers is a clear and resounding "no" IMHO.
You are entirely right when you say that real world programming tasks are often not quite so challenging. However, the ratio of people we need to solve boring real world programming tasks versus computer scientists we need to solve hard and interesting computationnal problems is extremely high. We really don't need that many computer scientist in the first place, yet we push every prospective candidate toward that discipline because most employers insist on a CS degree even for the lowliest entry level position in IT. This is ridiculous, and a waste of ressources too.
If you want to understand regular expressions, you've got to understand the mathematics of finite state automata.
That's a a pretty good demonstration, by the absurd, of what is wrong with considering CS as a good preparation for programming. No, you do not need to understand the mathematics of finite state automata to make good use of regular expression. I don't. I am pretty damn certain that none of my colleagues does either. Yet, we make use of regex daily. Maybe if I had to write my own regex engine I would need it. But it's a solved problem, so I just use the tools that are availables (and they're is plenty).
I don't want to flame, but I am 90% certain that you are either a student, or working in academics.
And what are you expecting to do as a job ? Programming ? Does your curriculum have prepared you well for this job ? You know, programmers in the real world do not have to write code that merge sort some files very often.
I don't mean to flame, but you just sound like the typical inexperienced CS graduate. Good luck anyway !
The problem with your position is that the job market have come to expect CS degree from programmers. Vocationnal school degrees, while being more down-to-earth in their approach, are considered inferior by the vast majority of employers. So, if you want to make a living programming, CS degree is considered a must. I know it make no sense in the real world, but would you rather get to begin your career on a lower rung because you choose the most appropriate curriculum for your career path ? For most people, the answer is no, thus we are collectively trying to retrofit computer "scientists" into programming roles.
Some people sugggested MIS as a better academic path for programmer. I don't know. At my University, the MIS curriculum involve a lot of business bullshit such as marketing or finance. I know these are good to know from the organizational point-of-view, but if you expect to produce decent programmers, you need to keep some focus.
Actually, I think there is no good path for those who want to get into programming in the current academic model. It's even worse for IT. What would a prospective system administrator take as degree ?
On one setup that has both sendmail and postfix, I know postfix loses far more mail than sendmail (which has lost none).
That smell a lot like FUD to me. Hopefully, you have been filling bug report to Postfix about it. Losing a single email is a serious bug, and I am pretty certain it would have been acted upon in the earnest.
Personnally, I never lost a single mail (that I know of) with Postfix. But I guess my anecdotes are no more worthy than yours.
I've spent time and learned how to make it do very unusual things (at the.cf macro level) and it is very powerful since it has a full programming language built in.
That's basically the only reason left to use sendmail : you have spent time learning it and you hope to get back on that investment. Some misplaced nostalagia that should be written off IMHO. In 2005, there is no reason for a fledging admin to go through the hardship of learning the sendmail mess, except if he have to support legacy installation.
And maybe I lack imagination, but I can't think of anything that could not be done in Postfix, given a content filter in last resort. What's the point of learning a spcialized mini-langage when you could use one that you already know, and that is more generally useful anyway ? And Exim is apparently even more programmable.
Preaching about sendmail programmability sound to me as a case of every problem looking like a nail to a hammer holder.
And precisely what make it the "best tools for the job" except a warm fuzzy feeling about it ?
I built a mail system for 100K accounts on Postfix. I have absolutely no doubt it would scale to 1M given appropriate hardware.
Anecdote : to make Postfix deliver mail via LMTP, you just add a "mailbox_command = lmtp:unix:/path/to/socket" directive (or a transport table entry saying approximately the same). Easy as pie. I once watched an experienced sendmail admin (think > 10 years experience) fight for over 15 minutes compiling m4 macro to achieve the same result with sendmail. I know nothing of sendmail except the very basic, and this incident taught it was not worth it for me to waste my time on it either.
Sendmail should be dead and buried. It's time have passed. Both Exim and Postfix are modern and well-maintained MTA that have been build with a clear idea of *not* to do, given sendmail history. And succeed pretty well at it too.
Really, the case is quite simple. If you charge for WiFi Internet access, I will not patronize your business. Why would I pay for a service provided free elsewhere ?
Slashdot Mirror
K3B and Klipper kept me on KDE for a looooong time.
There is no such thing as /usr/share/bin.
Point 1 and 2 could be solved by using by using Novell Linux Desktop. Yes, it's boring, but you can't have it both way.
BTW, the menu in my 9.3 installation have a single volume control applet, and it's pretty damn simple to understand. Actually, it's pretty much a copy of the Windows mixer. WTF is wrong with your install ?
Ok. But this does not tell me how hardware accelerated XML presentation layer "will change the computing world". Lot's of hype, very little substance.
You've read it on Slashdot first, folks : hardware accelerated XML presentation layer will change the computing world ! Forget quantum computing, you really more eye candies in the end.
I'm so excited, I just pissed on the carpet.
Ok, since there is not much substance in this thread, let's argue semantic ? Sorry, I am going this pass this turn.
Does such a thing exist ? Is there, somewhere, a software that I can buy that will satisfy all my needs in IT ?
Because, of course, the server room is not meaningful to "enterprise software".
FLOSS is, all thing being equal, cheaper and more flexible. So it's not a "zero-sum equation". The specific of your situation might call for a proprietary solution; I won't hold grudge to anybody in such a situation.
I don't understand why you are mixing enterprise and end-users software in the discussion, unless you confuse a word-processor for an "enterprise software". It's two different market entirely, and FLOSS serve the former much better than it does the latter.
You do a lot of broad generalization (FLOSS devs are elitist, FLOSS are crap, etc) but do not get much into the specifics. Suitability to task is a case-by-case issue. I do not think anybody in his right mind ever suggested you use vi as a word processor.
Lots of closed-source commercial fail these requirements too.
Enterprise-level application require professionnal to install and support them. That's a given. Why would it be any different with FLOSS ?
This is impossible with commercial software, except if the supplier is very small or you are a *very* large customer. Otherwise, you never get to make your suggestions to the developpers. You are always shielded from the production by the marketing and sales departement. You might tell your sales rep your suggestions, but what he does with them is impossible to tell. At least, with FLOSS, you have some measure of access to the devs. Wheter or not they will listen to your suggestion is not a given, but if you bring something worthwhile to the table, they might.
Of course, you have to work within their workflow (ie submit your request through the bug tracking system). At least, you know where your suggestions are going.
Why IBM or SuSE ? Why not hiring a local service firm ? Or developping in-house expertise ?
The nature of FOSS is not to provide "enterprise-level services"; it is to provide software. Now, if you want to buy "enterprise-level services" for FOSS, there is plenty of people willing to sell that you.
...
Again, what is the problem with FOSS ? That it does not come bundled with "enterprise-level services" for free ? Cry me a river
Both RedHat and Novell would come to you with marketable products. Remember, if you want shrink-wrap solution, you need to shop for shrink-wrap solution, and Sourceforge do not sell shrink-wrap solution (where RedHat and Novell do). But don't cringe about the price tag, because all that wrapping is costing mhtem oney, and this cost is being passed to you, the customer. Not that there is anything wrong with it, if you already pay for closed-source commercial software.
In the meantime, I make a living installing and supporting the applications you snob, and it cost my clients a fraction of the price of shrink-wrap software. Go figure.
Well, it's useful to computer scientists. It's peripherically useful to computer programmers, but these could be taught a more pertinent curriculum. IMHO. A curriculum that could include stuff like project management, interface design, software testing and, of course, more software engineering.
While I disagree with your point of view regarding the programmer's expected skillset, I must salute your position on drug screen test. Up here in Canada, (almost) nobody ask for screen test. It never happenned to me, but I think the practice is revolting. People must stand up. Otherwise, what next ? DNA sample ?
I agree 100% with everything you just wrote. My problem is not with the challenge level or the usefulness of CS, it is with the expectative that employers (and, to a certain degree, students) have with regard to CS curriculum. If we expect CS graduate to work as programmers, as it is most certainly the case right now, are we preparing them adequately ? The answers is a clear and resounding "no" IMHO.
You are entirely right when you say that real world programming tasks are often not quite so challenging. However, the ratio of people we need to solve boring real world programming tasks versus computer scientists we need to solve hard and interesting computationnal problems is extremely high. We really don't need that many computer scientist in the first place, yet we push every prospective candidate toward that discipline because most employers insist on a CS degree even for the lowliest entry level position in IT. This is ridiculous, and a waste of ressources too.
That's a a pretty good demonstration, by the absurd, of what is wrong with considering CS as a good preparation for programming. No, you do not need to understand the mathematics of finite state automata to make good use of regular expression. I don't. I am pretty damn certain that none of my colleagues does either. Yet, we make use of regex daily. Maybe if I had to write my own regex engine I would need it. But it's a solved problem, so I just use the tools that are availables (and they're is plenty).
I don't want to flame, but I am 90% certain that you are either a student, or working in academics.
And what are you expecting to do as a job ? Programming ? Does your curriculum have prepared you well for this job ? You know, programmers in the real world do not have to write code that merge sort some files very often.
I don't mean to flame, but you just sound like the typical inexperienced CS graduate. Good luck anyway !
The problem with your position is that the job market have come to expect CS degree from programmers. Vocationnal school degrees, while being more down-to-earth in their approach, are considered inferior by the vast majority of employers. So, if you want to make a living programming, CS degree is considered a must. I know it make no sense in the real world, but would you rather get to begin your career on a lower rung because you choose the most appropriate curriculum for your career path ? For most people, the answer is no, thus we are collectively trying to retrofit computer "scientists" into programming roles.
Some people sugggested MIS as a better academic path for programmer. I don't know. At my University, the MIS curriculum involve a lot of business bullshit such as marketing or finance. I know these are good to know from the organizational point-of-view, but if you expect to produce decent programmers, you need to keep some focus.
Actually, I think there is no good path for those who want to get into programming in the current academic model. It's even worse for IT. What would a prospective system administrator take as degree ?
That smell a lot like FUD to me. Hopefully, you have been filling bug report to Postfix about it. Losing a single email is a serious bug, and I am pretty certain it would have been acted upon in the earnest.
Personnally, I never lost a single mail (that I know of) with Postfix. But I guess my anecdotes are no more worthy than yours.
That's basically the only reason left to use sendmail : you have spent time learning it and you hope to get back on that investment. Some misplaced nostalagia that should be written off IMHO. In 2005, there is no reason for a fledging admin to go through the hardship of learning the sendmail mess, except if he have to support legacy installation.
And maybe I lack imagination, but I can't think of anything that could not be done in Postfix, given a content filter in last resort. What's the point of learning a spcialized mini-langage when you could use one that you already know, and that is more generally useful anyway ? And Exim is apparently even more programmable.
Preaching about sendmail programmability sound to me as a case of every problem looking like a nail to a hammer holder.
It's been done already; it's called "Postfix".
And precisely what make it the "best tools for the job" except a warm fuzzy feeling about it ?
I built a mail system for 100K accounts on Postfix. I have absolutely no doubt it would scale to 1M given appropriate hardware.
Anecdote : to make Postfix deliver mail via LMTP, you just add a "mailbox_command = lmtp:unix:/path/to/socket" directive (or a transport table entry saying approximately the same). Easy as pie. I once watched an experienced sendmail admin (think > 10 years experience) fight for over 15 minutes compiling m4 macro to achieve the same result with sendmail. I know nothing of sendmail except the very basic, and this incident taught it was not worth it for me to waste my time on it either.
Sendmail should be dead and buried. It's time have passed. Both Exim and Postfix are modern and well-maintained MTA that have been build with a clear idea of *not* to do, given sendmail history. And succeed pretty well at it too.
You see, that's part of the problem. MIS is still too focused on administration. Marketing and finance courses ? Now that's useful to an IT worker !
So, between CS and MIS, there is no path to adequately prepare IT workers for the real world in current University curriculum.
Well, it suck because SSHA is the strongest password hash available in OpenLDAP, according to slapd.conf(5).
So tell me : should I stop using SSHA to store password hash ?
RaLink, Atheros and Centrino chipsets are well-supported in Linux. There is probably more I can't think of too.
Really, the case is quite simple. If you charge for WiFi Internet access, I will not patronize your business. Why would I pay for a service provided free elsewhere ?