Why not try Zsh? It has all three forms of history, !-history, fcedit and arrows/I-search. Most of which can be turned on and off at your leisure. And don't forget the many other improvements over most other shells in use today.
I wrote an article for daemon news some time ago about the advantages of zsh over bash. Check it out.
The E10K (starfire) isn't a patch on the IBM mainframes that the article talks about. You can only configure domains based around system boards, and you need to use an individual disk to install a separate copy of the OS for each domain. On top of this, I have found that dynamic reconfiguration isn't as reliable as I'd like it to be. OTOH, we are using quite an old starfire now.
You can get large amounts of storage and redundancy on a starfire. But it still isn't a mainframe, although it is still a damn good high end Unix server.
I totally agree. The whole point is not about packages; we have plenty of types of these already. It's about *meta* packages, and creating these from source distributions so that they can be easily integrated into your operating system.
eg: If I compile a program on my Solaris box, I want a pkg for it to be created in pkgadd format, so that I can distribute it to my other Solaris boxes. Likewise, on FreeBSD, I want it in pkg_create format. And on Linux, it needs to be RPM.
What is needed is a package creator that understands different Operating Systems styles of packages. Dammit, if I compile under Windows, it should be able to create a installshield package for me!
Note that since Solaris 7, you get UFS logging for free with the base system. If you have the server edition of the previous operating systems, then you get Disksuite which can do the same. This should reduce fsck times significantly, although (according to some) it's not *true* journalling.
The reason that BSD development appears to be on a par with Linux is simply that there are a limited number of people capable of coding at the level required. Whilst Linux (probably) has more hard-core kernel coders overall, as a percentage figure above the number of people doing similiar things to BSD, it might not be such a great increase.
And then you come to some point in a system which is horrifically complicated (such as the VM system) where you probably have under 10 people on each side that understand it thoroughly.
On the other hand, practically anybody with a little bit of C knowledge and patience can whip up a nice looking GTK+ application without too many problems, so here, the larger user base of Linux shows up as more user applications are written.
Note that all this is just theorizing; I have no figures to back me up.
I've worked with quite a few of the above and I don't like them in the way that you do. I'm sure they're good products when used correctly, but they have a *huge* learning curve when compared to the average piece of open source software. And quite frequently, their documentation sucks rocks too. Also, as I mentioned earlier, these products usually don't integrate very well at all with their surrounding environment.
So, you end up spending thousands of currency units more on consultants and/or training. And usually, the consultants leave behind little or no documentation on what they've done, so they have to be called back if you want to make any changes at all.
I for one would *gladly welcome* some commercial software that didn't suck in this manner.
No, you should not run towards NetBackup. See my earlier comment attached to the first thread about why NetBackup is a completely foetid piece of software death.
Disclaimer: I do like some veritas software including Volume Manager and VxFS.
Veritas NetBackup is *crap*. It's an old product made by a company called OpenView in the 80's which has never been properly integrated with modern Unix. The interface is a nightmare, it installs 3 tons of shit all over your filesystem in places you don't want it. It installs 3 dozen new services into inetd (a service for each tape drive ro something like that!!!). The GUI is woefully inconsistent and counter intuitive. Overall, I find it very inflexible and difficult to automate. On top of this, the documentation is pathetic. Oh, and don't go on the training course, you just get to sit in front of a power point presentation for three days. And guess what the course documentation is? Copies of the powerpoint slides. *bangs head against desk*.
I have looked at this several times over the past year and each time, I have gone back to plain old gnutar.
Also, I note that they are violating the GPL by including an old version of GNU Tar (1.09, so old it doesn't use --help, but +help instead) with modifications, but no source code available.
It's a shame, but I find this symptomatic of practically every piece of commercial Unix software that I have seen. They treat the system as something to be pushed away and ignored, rather than integrated with. So, I invariable end up going with the open-source alternative if at all possible. Ah well, their loss of sales.
To be fair, I think that the software would work a lot better if I had a tape library available to me instead of the poxy DLT4700 stacker that I have now. You may be able to get a better (and less biased) opinion than mine over at backupcentral.com.
Solaris simply doesn't have the momentum behind it on x86 that FreeBSD does. I'm not a fanatic, but I certainly don't think that just because Solaris 8 is available for free with source pseudo-available everyone working on FreeBSD will suddenly just stop and go "Shucks, guys, this is what I was after all along". Hell, no. There may well be some cross-breeding of ideas between the two, but it's more likely to be that FreeBSD includes stuff from Solaris than the other way around.
And also, think how likely it is that Sun will actually implement the changes in Solaris that *you* want. Even/if/ you submit patches. And how often do you think they'll release updates? Your chances of getting something even remotely close to CVSup are minimal.
Solaris 8 supposedly has a lot of GNU tools (including the ones mentioned above). They're finally getting a clue it looks like....
Yes, so they're finally going to ship Perl with Solaris 8. Big fscking deal. And how regularly do you think they'll update all these utilities they are suddenly including? You'll be stuck on whatever ancient version of perl they choose to throw in. And of course, the one utility that you want to be in there won't be. Oh, and if they do decide to update them, the update will only be available to contract customers. In short, whilst this is nice for some customers, for the majority it's not what they need. They need something like FreeBSD's ports or RPMs (although Solaris already has it's own packaging scheme, so the latter is less useful).
After a long period of administering sun boxes, I have come to the conclusion that I don't trust sun to get it right. I have a hundreds-of-lines-long script just to fix up their boxes after the install...
Their crass behaviour is vindicated by their obsession with the unhealthy SCSL and now this. I won't be sorry when they do a U-turn on Linux and many other matters later this year.
A very lucid explanation of a complicated system. Like many good Unix people, Matt Dillon writes natively as well as he does in code. I only hope that this encourages me to take up poking around in FreeBSD's kernel code...
I have to vote for zsh. It sure doesn't look as pretty as the others mentioned, but it's the fastest and most efficient way of interacting with practically all parts of the known Unix system. Once you've tried it, you won't want to go back to anything else.
Well, for me, the biggy is full IPV6 support. However, you can view the release notes in CVS already, to get an idea of the feature list that will be posted with the release. Point your browser at RELNOTES. TXT, and click on the latest revison number (1.54 as of writing).
Funny indeed. I've never managed to install *anything* without a hitch, except for FreeBSD (which I can practically do blindfolded now).
The real question, though is how easy it is to stay up to date once it's installed, since you're likely to be doing that much more often than actually installing/reinstalling (if you're not a windows user, anyway). Again, I find FreeBSD's cvsup && make world mechanism to be as near hands off as one can get. Simply beautiful.
Re:Agreed. On FreeBSD is crashes on almost every p
on
Mozilla Status Update
·
· Score: 1
If you use the port in/usr/ports/www/mozilla, it's much more stable. Although it does have trouble with slashdot, most other things work quite well.
It's not. Office 2000 comes in at over a hundred megs for a minimal install. My latest XEmacs install (with the sumo tarball of all packages) was around 70Mb. And you can strip it down a *lot* further if you want.
I know it's probably a bit late, but if you'd have chosen FreeBSD, the license would have allowed you to withhold source code as much as you wish. A lot of other "Appliance" companies are doing just this, for instance http://www.whistle.com/ and the GNATbox firewall (forgot URL).
When I worked at Demon, the netapps were one of the most reliable pieces of machinery that I administered. Whilst you might think that network attached storage can be a performance problem, in practice it worked very well indeed.
You do, however, need to be aware of how to make your application play well over NFS. Exim is actually reasonable at this. Qmail is good at storing mailboxes on NFS thanks to it's Maildir technology, but the mail queue *needs* to be on a local disk... I'm not sure about postfix or sendmail (bletch).
Unfortunately, I can't remember the command to make the individual LEDs on the disks blink, which is one of the best remote diagnostic features ever.:-)
This is a subject coming up more and more frequently on slashdot. I think it's time that we contacted somebody at the patent office for the weekly interview. You never know, it might make a difference. At the very least, they might tell us how we can make a difference.
This is where the centralised method of distribution that FreeBSD et al use really wins. You just set up CVSup to run regularly and run "make world" when you need to actually install the patches. Strictly a hands off operation.
Slashdot Mirror
Why not try Zsh? It has all three forms of history, !-history, fcedit and arrows/I-search. Most of which can be turned on and off at your leisure. And don't forget the many other improvements over most other shells in use today.
I wrote an article for daemon news some time ago about the advantages of zsh over bash. Check it out.
The E10K (starfire) isn't a patch on the IBM mainframes that the article talks about. You can only configure domains based around system boards, and you need to use an individual disk to install a separate copy of the OS for each domain. On top of this, I have found that dynamic reconfiguration isn't as reliable as I'd like it to be. OTOH, we are using quite an old starfire now.
You can get large amounts of storage and redundancy on a starfire. But it still isn't a mainframe, although it is still a damn good high end Unix server.
I totally agree. The whole point is not about packages; we have plenty of types of these already. It's about *meta* packages, and creating these from source distributions so that they can be easily integrated into your operating system.
eg: If I compile a program on my Solaris box, I want a pkg for it to be created in pkgadd format, so that I can distribute it to my other Solaris boxes. Likewise, on FreeBSD, I want it in pkg_create format. And on Linux, it needs to be RPM.
What is needed is a package creator that understands different Operating Systems styles of packages. Dammit, if I compile under Windows, it should be able to create a installshield package for me!
-Dom
And this will install all the dependencies, as well as downloading everything from the Internet. Just come back later.
Note that since Solaris 7, you get UFS logging for free with the base system. If you have the server edition of the previous operating systems, then you get Disksuite which can do the same. This should reduce fsck times significantly, although (according to some) it's not *true* journalling.
The reason that BSD development appears to be on a par with Linux is simply that there are a limited number of people capable of coding at the level required. Whilst Linux (probably) has more hard-core kernel coders overall, as a percentage figure above the number of people doing similiar things to BSD, it might not be such a great increase.
And then you come to some point in a system which is horrifically complicated (such as the VM system) where you probably have under 10 people on each side that understand it thoroughly.
On the other hand, practically anybody with a little bit of C knowledge and patience can whip up a nice looking GTK+ application without too many problems, so here, the larger user base of Linux shows up as more user applications are written.
Note that all this is just theorizing; I have no figures to back me up.
I've worked with quite a few of the above and I don't like them in the way that you do. I'm sure they're good products when used correctly, but they have a *huge* learning curve when compared to the average piece of open source software. And quite frequently, their documentation sucks rocks too. Also, as I mentioned earlier, these products usually don't integrate very well at all with their surrounding environment.
So, you end up spending thousands of currency units more on consultants and/or training. And usually, the consultants leave behind little or no documentation on what they've done, so they have to be called back if you want to make any changes at all.
I for one would *gladly welcome* some commercial software that didn't suck in this manner.
No, you should not run towards NetBackup. See my earlier comment attached to the first thread about why NetBackup is a completely foetid piece of software death.
-Dom
Disclaimer: I do like some veritas software including Volume Manager and VxFS.
Veritas NetBackup is *crap*. It's an old product made by a company called OpenView in the 80's which has never been properly integrated with modern Unix. The interface is a nightmare, it installs 3 tons of shit all over your filesystem in places you don't want it. It installs 3 dozen new services into inetd (a service for each tape drive ro something like that!!!). The GUI is woefully inconsistent and counter intuitive. Overall, I find it very inflexible and difficult to automate. On top of this, the documentation is pathetic. Oh, and don't go on the training course, you just get to sit in front of a power point presentation for three days. And guess what the course documentation is? Copies of the powerpoint slides. *bangs head against desk*.
I have looked at this several times over the past year and each time, I have gone back to plain old gnutar.
Also, I note that they are violating the GPL by including an old version of GNU Tar (1.09, so old it doesn't use --help, but +help instead) with modifications, but no source code available.
It's a shame, but I find this symptomatic of practically every piece of commercial Unix software that I have seen. They treat the system as something to be pushed away and ignored, rather than integrated with. So, I invariable end up going with the open-source alternative if at all possible. Ah well, their loss of sales.
To be fair, I think that the software would work a lot better if I had a tape library available to me instead of the poxy DLT4700 stacker that I have now. You may be able to get a better (and less biased) opinion than mine over at backupcentral.com.
Solaris simply doesn't have the momentum behind it on x86 that FreeBSD does. I'm not a fanatic, but I certainly don't think that just because Solaris 8 is available for free with source pseudo-available everyone working on FreeBSD will suddenly just stop and go "Shucks, guys, this is what I was after all along". Hell, no. There may well be some cross-breeding of ideas between the two, but it's more likely to be that FreeBSD includes stuff from Solaris than the other way around.
/if/ you submit patches. And how often do you think they'll release updates? Your chances of getting something even remotely close to CVSup are minimal.
And also, think how likely it is that Sun will actually implement the changes in Solaris that *you* want. Even
Just because it's free doesn't mean it's good.
Yes, so they're finally going to ship Perl with Solaris 8. Big fscking deal. And how regularly do you think they'll update all these utilities they are suddenly including? You'll be stuck on whatever ancient version of perl they choose to throw in. And of course, the one utility that you want to be in there won't be. Oh, and if they do decide to update them, the update will only be available to contract customers. In short, whilst this is nice for some customers, for the majority it's not what they need. They need something like FreeBSD's ports or RPMs (although Solaris already has it's own packaging scheme, so the latter is less useful).
After a long period of administering sun boxes, I have come to the conclusion that I don't trust sun to get it right. I have a hundreds-of-lines-long script just to fix up their boxes after the install...
Their crass behaviour is vindicated by their obsession with the unhealthy SCSL and now this. I won't be sorry when they do a U-turn on Linux and many other matters later this year.
A very lucid explanation of a complicated system. Like many good Unix people, Matt Dillon writes natively as well as he does in code. I only hope that this encourages me to take up poking around in FreeBSD's kernel code...
I have to vote for zsh. It sure doesn't look as pretty as the others mentioned, but it's the fastest and most efficient way of interacting with practically all parts of the known Unix system. Once you've tried it, you won't want to go back to anything else.
-Dom
Funny indeed. I've never managed to install *anything* without a hitch, except for FreeBSD (which I can practically do blindfolded now).
The real question, though is how easy it is to stay up to date once it's installed, since you're likely to be doing that much more often than actually installing/reinstalling (if you're not a windows user, anyway). Again, I find FreeBSD's cvsup && make world mechanism to be as near hands off as one can get. Simply beautiful.
If you use the port in /usr/ports/www/mozilla, it's much more stable. Although it does have trouble with slashdot, most other things work quite well.
-Dom
Is that why Microsoft are reupted to be bringing out Stake 2000, which has even less wood and is blunter than ever?
-Dom
It's not. Office 2000 comes in at over a hundred megs for a minimal install. My latest XEmacs install (with the sumo tarball of all packages) was around 70Mb. And you can strip it down a *lot* further if you want.
:-)
Plus it's faster than Word.
You are Ian Hislop and I claim my free Paul Merton!
:-)
I know it's probably a bit late, but if you'd have chosen FreeBSD, the license would have allowed you to withhold source code as much as you wish. A lot of other "Appliance" companies are doing just this, for instance http://www.whistle.com/ and the GNATbox firewall (forgot URL).
Good luck, anyway!
When I worked at Demon, the netapps were one of the most reliable pieces of machinery that I administered. Whilst you might think that network attached storage can be a performance problem, in practice it worked very well indeed.
:-)
You do, however, need to be aware of how to make your application play well over NFS. Exim is actually reasonable at this. Qmail is good at storing mailboxes on NFS thanks to it's Maildir technology, but the mail queue *needs* to be on a local disk... I'm not sure about postfix or sendmail (bletch).
Unfortunately, I can't remember the command to make the individual LEDs on the disks blink, which is one of the best remote diagnostic features ever.
-Dom
This is a subject coming up more and more frequently on slashdot. I think it's time that we contacted somebody at the patent office for the weekly interview. You never know, it might make a difference. At the very least, they might tell us how we can make a difference.
-Dom
Grow up. There's plenty of room for both. As I think you'll agree by looking at Microsoft, a one-pony show is in nobody's interest.
-Dom
Don't forget all the work that they sponsored on XFree86. They do contribute stuff back to the community.
This is where the centralised method of distribution that FreeBSD et al use really wins. You just set up CVSup to run regularly and run "make world" when you need to actually install the patches. Strictly a hands off operation.