Re: Advocacy thoughts - was Re: bad nasty evil thread
On Thu, Jan 24, 2002 at 04:00:41PM +, Jonathan Peterson wrote: > Not always, but it depends how hard you try... Yep... as anyone who's seen me on IRC over the last few days will testify, I'm struggling to get Debian working on my laptop. Not (just) out of some innate desire to rid myself of Windows, but because all the *nix administration experience at my company is on servers... I can whip up a BSD server install in a couple of hours but I've never tried to get X working. We've found the right motivations in our company - mostly financial. We're already deploying StarOffice on some of our Windows desktops to avoid the cost of more MS licenses. We've explained the advantages in having proper networked storage, hotdesking (through LDAP and remote-mounting /home), a simpler backup policy, easier remote administration, stuff like that... and they like the idea. Especially, they like the lack of licensing. We've demonstrated that people can do all they need to do, with open source software. Management are clueful enough to realise that people *will* bitch but get used to it - the plan is for IT to train one person in each department on the basics of using a Linux desktop and StarOffice and use them as first-line support. Get the users to teach the users, and they'll be more likely to grok it.. and with no root, they can't kill things too badly. In other areas, I've replaced propriaty FileMaker Pro databases with Postgres databases, interfaced with a web browser, using perl and TT. I've saved a fortune in licensing in a few months, and improved the system to boot. Most of my admin scripts are in perl, just because it's what I know... and they're nicely documented with comments and POD, which means I can create my IT procedures manual (now stored under CVS along with all my other content, a new introduction to my company) automagically. It does depend on the company, and the angle you find. In my case, it wasn't "We must be doing it this way" but "Windows works, here's an alternative that will save us time and money, why don't we give it a go?" If the end-user trial of Debian desktops doesn't go well, we'll stick with Windows, which is why I'm putting so much effort into learning Debian admin - we need everything to work perfectly, first time, so our users don't get that bad experience thing. Alex -- KCBpd lWmulvo ECS+ m5 CPEIV B13 Ou Lmb Sc+isIC+ T++ A6LAT H6oe b5 D+ - See http://bob.bob.bofh.org/~giolla/bobcode.html for decoding Website: http://www.cpio.org/~grimoire Writing: http://www.livejournal.com/~diffrentcolours
Re: Advocacy thoughts - was Re: bad nasty evil thread
On Thu, Jan 24, 2002 at 04:00:41PM +, Jonathan Peterson wrote: > serving platform*. And Penderel is hardly a testament to the reliability > of Linux machines. I think Penderel is suffering from a bad case of old-shabby-hardwaritis. Paul
Re: Advocacy thoughts - was Re: bad nasty evil thread
David Cantrell wrote: > > Does he have a friend > > who had real trouble setting Samba up for some reason and now he's wary > > of it? > > If he values the advice of a random friend over the advice of the person > he pays good money to to know about these things, then he's an idiot. No, actually I disagree, and it's an important point. I didn't say it was a random friend. My friends are less random than my employees. I employ people based on the best of a bunch of C.V.s that I manage to get within a short time period, minus those ruled out by salary restrictions, minus those that don't want the position in the end, and filtered according to one or maybe two hour long interviews and if I'm lucky some examples of past work. My friends on the other hand may be people I've worked with closely for years, selected out of all the many people I've ever worked with. Or people whose opinions I've read on mailing lists for years. Like many of us I tend to change jobs every 1-2 years, so I'd never know my employees longer than that, while I've known some of my technically minded friends for 5 years or more. I have two employees who work for me. They are both good(ish) Java programmers but have limited knowledge of other areas of computing. One of them tells me he wants to use some commercial bug tracking tool that he used in his last job. It's expensive and proprietary (PVCS tracker) but he really wants it. My friends tell me they've got by fine with cheap and chearful bits of web based freeware, or maybe bugzilla when things get hairy. I overule him and tell him we'll use some PHP based thing I installed in my lunch break (called Mantis, it's on sourceforge somewhere). Why should I do this? He's the developer who has to use it, right? I should just give him what he recommends, after all he says it's good and I've never even used it. Well, my job as manager is not to do whatever my programmers say. I have to think about budget, I have awareness of our future staffing requirements, which make the license fee look a lot worse down the line. In fact, I had a political aim too - I recognised this employee was rather over fond of commercial software, and I wanted to show him that sometimes free stuff can do the job - and it's much better to show than to tell. Plus, it wouldn't kill him to learn a bit of PHP, if nothing else so he can compare it to Java. Plus, if I'm wrong, it's not really a big deal to fix down the line, it's a good chance for me to try something out. In fact, I was also quite interested to see how he would react to being overuled by me. The point is, I did not evaluate this based purely on "What is the best (value for money) bug tracker around". I had lots of other things on the agenda. And I had a lot of faith in friends who assured me that for my size projects you really don't need super duper feature rich wizzo bug tracking tools. Business people place a huge amount of store in the opinions of their friends and contacts. Unless they have a good relationship with their employees, they may place considerably less store in their opinions. If you want your boss to listen to you, be nice to him. Yes, that's politics, but it's a feature of human nature so we have to deal with it. > > I've met a lot of resistance to Perl based on 'bad past experiences'. We > > all know these were usually caused by bad programmers and stupid > > timescales. But your bad experience counts for a lot more than someone > > else's good experience, which is why the 'Perl success stories' are not > > alway so convincing. > > That's OK then. Soon *everyone* will have been burnt by Java and come > flocking back :-) Actually I think that's true. Already Java is becoming the new VB. Want a job in computing? Can't do anything? Read a book on Java and tell them you are a programmer. As the quantity Matt Wright style Java and JSP increases (and I've seen plenty already) it will lose much of its appeal. I think Java will soon become what C is now. Neither good nor bad, still widely used, but somehow a bit old and tired, OK for some big old J2EE project left over from 2002 but not really the best thing to be doing a new project in. > One person can change the world, but most of the time they shouldn't > -- Marge Simpson Everyone has better sigs than me. -- Jonathan Peterson Technical Manager, Unified Ltd, +44 (0)20 7383 6092 [EMAIL PROTECTED]
Re: Advocacy thoughts - was Re: bad nasty evil thread
On Thu, Jan 24, 2002 at 04:00:41PM +, Jonathan Peterson wrote: > > > Anyone who chooses a fileserver which won't work reliably over one that > > will - > > Neither of you can predict the future. I don't know the specifics of > this case, but Win2K (or even NT3SP3) is hardly an unreliable file > serving platform*. And Penderel is hardly a testament to the reliability > of Linux machines. Penderel is a testament to the unreliability of x86 machines. It is my experience, however, that on any particluar x86 box, Linux is more stable than Billware. With Linux you only have to worry about the shitty hardware. With Billware, you need to worry both about that and about the shitty OS. > I doubt the pretty login screen counts for much, although possibly some. > Seriously, what do you suppose this manager's reasons were for making > his/her decision? Was it a simple case that he sees Microsoft as a > predictable if unspectacular option versus Linux as a high risk option? If he sees it as a high-risk option, then he is at best ignorant. > Does he just think you recommended Linux only because you can't be > bothered to learn how to administer NT properly? If he thinks Roger is unprofessional then he should say so, and should fire him. That he didn't shows that he is an idiot if that is his reason for choosing NT. > Does he have a friend > who had real trouble setting Samba up for some reason and now he's wary > of it? If he values the advice of a random friend over the advice of the person he pays good money to to know about these things, then he's an idiot. >Was it that the decision over this file server was entirely > trivial as far as he was concerned, and he just couldn't care less what > gets used? If that is the case, then he's an idiot. > Occasionally you do get incompetant people. But usually, people have > good reasons for making decisions, even when they make the wrong > decision. This seems to boil down to a pointy-hair not paying attention to the people who he pays to know about these things. Not paying attention to your hired experts whilst continuing to pay for the service of those hired experts is idiocy. >Until you understand why people decide to choose technically > inferior products, it's hard to make your technically superior product > more popular with them. Agreed. But I see no non-technical reasons above which would justify choosing something which is technically inferior. > I've met a lot of resistance to Perl based on 'bad past experiences'. We > all know these were usually caused by bad programmers and stupid > timescales. But your bad experience counts for a lot more than someone > else's good experience, which is why the 'Perl success stories' are not > alway so convincing. That's OK then. Soon *everyone* will have been burnt by Java and come flocking back :-) > [stuff about negative campaigning being bad] all true. -- David Cantrell | [EMAIL PROTECTED] | http://www.cantrell.org.uk/david One person can change the world, but most of the time they shouldn't -- Marge Simpson
RE: Advocacy thoughts - was Re: bad nasty evil thread
Since NT4SP3 is not even Y2K ready, I doubt that anyone would feel it is a fair comparison. I suppose you could compare it to Linux 2.0.*. My experience is that NT, Windows and Unix can be very stable platforms - they also can be unstable and I have seen it in all three systems. What about the bugs that are found in various kernel versions, my system at home can be crashed by USB printing until I moved to 2.4.10. Its a known bug but took about 4 months to fix. I'm not trying to single out one system over another, but they all have strengths and weaknesses. Richard > -Original Message- > From: Roger Burton West [SMTP:[EMAIL PROTECTED]] > Sent: 24 January 2002 16:16 > To: [EMAIL PROTECTED] > Subject: Re: Advocacy thoughts - was Re: bad nasty evil thread > > On Thu, Jan 24, 2002 at 04:00:41PM +, Jonathan Peterson wrote: > > >Neither of you can predict the future. I don't know the specifics of > >this case, but Win2K (or even NT3SP3) is hardly an unreliable file > >serving platform*. > > In my experience they both are (assuming you meant NT4SP3). > > >And Penderel is hardly a testament to the reliability > >of Linux machines. > > I don't know who set up Penderel. I do know that no Linux machine I've > set up has had downtime from anything other than hardware failure. > It's > my job to get these things right, and I do. > > I can't go into much more detail of this particular case, for obvious > reasons; suffice it to say that technical issues were not a factor. > > Roger > > -- > He's a suicidal guerilla barbarian whom everyone believes is mad. > She's > a virginal hypochondriac mercenary with an MBA from Harvard. They > fight > crime!
Re: Advocacy thoughts - was Re: bad nasty evil thread
On Thu, Jan 24, 2002 at 04:00:41PM +, Jonathan Peterson wrote: >Neither of you can predict the future. I don't know the specifics of >this case, but Win2K (or even NT3SP3) is hardly an unreliable file >serving platform*. In my experience they both are (assuming you meant NT4SP3). >And Penderel is hardly a testament to the reliability >of Linux machines. I don't know who set up Penderel. I do know that no Linux machine I've set up has had downtime from anything other than hardware failure. It's my job to get these things right, and I do. I can't go into much more detail of this particular case, for obvious reasons; suffice it to say that technical issues were not a factor. Roger -- He's a suicidal guerilla barbarian whom everyone believes is mad. She's a virginal hypochondriac mercenary with an MBA from Harvard. They fight crime!
Advocacy thoughts - was Re: bad nasty evil thread
> Anyone who chooses a fileserver which won't work reliably over one that > will - Neither of you can predict the future. I don't know the specifics of this case, but Win2K (or even NT3SP3) is hardly an unreliable file serving platform*. And Penderel is hardly a testament to the reliability of Linux machines. when the one that will has already been paid for, and the one > that won't will cost the company about L2,000 extra - simply because the > unreliable one has a pretty login screen _is_ an idiot. I'm sorry; I doubt the pretty login screen counts for much, although possibly some. Seriously, what do you suppose this manager's reasons were for making his/her decision? Was it a simple case that he sees Microsoft as a predictable if unspectacular option versus Linux as a high risk option? Does he just think you recommended Linux only because you can't be bothered to learn how to administer NT properly? Does he have a friend who had real trouble setting Samba up for some reason and now he's wary of it? Was it that the decision over this file server was entirely trivial as far as he was concerned, and he just couldn't care less what gets used? Occasionally you do get incompetant people. But usually, people have good reasons for making decisions, even when they make the wrong decision. Until you understand why people decide to choose technically inferior products, it's hard to make your technically superior product more popular with them. I've met a lot of resistance to Perl based on 'bad past experiences'. We all know these were usually caused by bad programmers and stupid timescales. But your bad experience counts for a lot more than someone else's good experience, which is why the 'Perl success stories' are not alway so convincing. In the past I've had surprising advocacy success using Perl for some very simple task, that any language would have been fine for. Just showing that a bit of Perl can be in your system without causing a meltdown can do a lot to re-assure management. Despite Microsoft's infamous FUD tactics, negative campaigning is not very effective. If people have actually had bad experiences with technology X, then convincing them to use something else is hardly a challenge. Not much point telling them what they already know. If on the other hand they have had good, or even just 'ok' experiences, then it's hard to convince them not to believe their own eyes. It's much easier to convince them that, sure technology X is pretty good, but not _as_ good as technology Y!!. Compare: Geek - "NT?? Don't use that, it's crap, it crashes all the time, you're stuck buying MS products forever, I mean every _technical_ person knows how much it sucks!" Manager - "Err, well that little file server seems to have been OK for the last two years. Sometimes on Monday morning it seems to go really slow - I just reboot it and it's fine, no big deal. I don't think we've had to buy lots of software we didn't expect to. I know the anti virus software doesn't come from Microsoft and it seems to work OK. I dunno, might as well stick with it". With: Geek - "NT?? That's not bad for simple file serving and stuff. But probably one of the things you've found is that you're constantly updating the data files for that anti virus package? That's kind of a pain to do all the time, and Linux doesn't really have a problem with viruses in the same way. Another thing you've probably noticed is how it can be a pain to move lots of files around at once, like when you want to move all the old purchase orders to that archive folder in the Finance share? Well, on Linux there are much better tools for that, kind of like those batch files you've got but really much more sophisticated. Those can help organise stuff alot, so you don't get cluttered up with old versions of files everywhere." Manager - "That's pretty interesting, but it's going to be a pain to move all the old stuff over, right?" Geek - "Yeah, a bit but not too much. The key thing is that the users can't tell the difference - both NT and Linux with Samba look the same to Windows desktops, so you won't have to retrain any-one - won't even have to get people to update their shortcuts or anything!" > there's no way round that. Not always, but it depends how hard you try... Jon * At this point someone comes up with a story about how they once had this NT machine that crashed every day blah blah and all it was doing was printing one word document a month blah blah blah and then they put Linux on it and it never crashed ever blah blah blah. Since these stories are not reflected by _any_ of my experiences in the matter, I'm unpursuaded by them. -- Jonathan Peterson Technical Manager, Unified Ltd, +44 (0)20 7383 6092 [EMAIL PROTECTED]
