Slashdot Mirror
Ease of Use vs. Sweat Equity
Kelly McNeill pointed us over to a new article: "Browsing through ZDNet's feedback a while back, I lighted on the now standard debate on the merits of Linux and related Operating Systems versus those of Microsoft Windows NT. One thing eventually got me thinking. In every posting that claimed success with using NT, the factor that was claimed as guaranteeing success, was never NT's 'superior' technology, which has been Microsoft's line all along, but instead the time spent in doing things right. In other words, the reason why some companies could claim success with their deployment of NT was good old elbow-grease/sweat equity. "
I'm the sysadmin in a cybercafé, and consequently I regularly have to explain the use of the computers to people who've never used a computer before in their lives. This explanation can even be as basic as 'watch the arrow on the screen. when you move the mouse up...'
All the public machine use W9x (people need games and MS Word), and all the servers are Linux. I'm looking at putting a few Linux machines out for the masses to use becuase if you've never used a computer before Windows isn't easy to learn The web integration in 98 just makes things worse.
To be fair, Linux is almost as difficult. But for the absolute neophyte, I'd rather spend time showing them how to use Linux than have to go back time and again to explain what happens when the screen goes all blue.
From what we have found just removing the necessity to use a mouse to run any application, from database management tools, to FTP clients is enough to slow us down such that we are faster under linux. Not to mention that we can code a web app in perl with mysql way faster than VBScript and SQL, that may be because we are more used to it, but somehow I doubt that. Not to mention the fact that we are more used to it, because it was easier to learn 3 years ago (and was around for cheap) than to pick up related MS technology. As to the "battle" between M$ and Linux, I think that high bandwidth is the only weapon that will effectively make Linux succeed. As high bandwidth becomes very widely accessible (no pun intended) people will want to run their own ftp servers, mail servers, web servers, dns, etc. This will increase the call for stable, reliable services at home that run constantly. ANd combined with this need is the low cost necessity. Why pay $5000 for hardware and software (IIS, SQL, NT, etc.) when you can get it all free for Linux, chalk up a $1000 server system and get ready to roll.......We need high bandwidth! (not to mention I wnat to be able to download an mp3 in less than 45 minutes at home :-)
Funny and I thought Perl == Paid employment recently located
- On the up side, that scares away the ludicrously incompetent
- Also on the up side, the expectation of difficulty/complexity means that people expect there to be some difficulty in figuring things out.
In contrast, those that expect hugs and kisses and simplicity because they're deploying NT run into opposite problems:When the situation (e.g. - independent of the OS in place) happens to be difficult/complex, this then doesn't phaze anybody, as they were prepared for there to be some difficulty.
As for the author's efforts at writing science fiction, it sounds like a case where you hope many of his neighbours are MSFT-critters, so that if his characters come for a meal, few will feel worried about it...
If you're not part of the solution, you're part of the precipitate.
I agree. Any OS can be -made- to do what you want, with enough time and effort. The question is not what you can do (which, according to Turing, was anything you liked that was computable), but how practical it is, and how much effort is involved.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
It seems to me that there are 2 kinds of company, and 2 types of IT depts.
Managers make the decisions about hardware/software and other business decisions and very often they are not in the 'know' about the arguments between opensource and closed source. They see a product that does 80% of what they want, and they will buy it because they *know* it will do what it says on the tin. They expect no more.
This is a standard thing, unless managers relinquish at least some of the responsibility to those in the 'know', then it's the way it stays.
To a non technical manager, paying $50,000 for a system that will do exactly what it says for minimal hassle then it's very tempting. Much more of a sure fire thing than spending half that on something which has the potential, but needs developing.
NT does to a greater or lesser extent do exactly what it says, as do the products that run on it. No they aren't great, and you can't always adapt them at all *because* they are proprietary, but most people (read managers) will settle for this.
That's why NT is here to stay.
The term ease of use is something of a misnomer when applied to WinNT, what it should be is ease of learning. WinNT has a much lower learning curve than that of Unix, but once you reach a certain level of proficiency in both, you will actually find that many tasks are far easier under a unix system.
;)
The problem is, of course, remembering the right incantations and understanding what they actually do, rather than clicking a few buttons, selecting some radio-buttons and then rebooting.
What it comes down to really, is what you class as 'ease of use'. I find Unix much more natural to use than NT, but I'm a bit of a masochist when it comes to computers -- if it doesn't hurt you aren't doing it right
--
I quite agree with the conclusion that successful NT implementations rely on elbow grease. I've done a few in my time.
And I've seen the way that management at big companies work. Once thay've decided on a "project" they pour resources into it like there's no tomorrow. You know, hire a few more NT sysadmins for good measure. Deploy random product X to a thousand machines because it's the flavour of the month.
Linux needs to get into the position where the same amount of resources get thrown at it. This is a big task, and it needs a concerted effort on the part of the open source community. Most corporations have a "better the devil you know" mentality, so some serious PR work needs to be done.
My number one hint:
Get a consortium of Linux consultants to put together a "showcase" 100% Linux enterprise system. Linux servers, email, e-commerce, intranet, security and desktops. Put up a top quality website showing how it was all done, the hardware chosen, the software used and the configuration steps taken.
This will give Linux some serious credibility. As a bonus, the partners in the venture will make a mint by helping out with implementations for companies that want a similar system. And it will show the commercial viability of supporting Linux to everyone else.
Authors background:- --------------------------------
-----------------------------------------------
Wesley Parish is a Linux user, and also plays around with Minix. He once tried to install a beta of Windows NT 3.51, but encountered a BSOD. He is still working on his anthropological SF novel, and has decided he would never invite his characters around for a meal, as they would eat the neighbors' dogs. Not such a bad idea, but if his neighbors came over to complain, his characters would eat his neighbors...
Translation: Wesley Parish is a college student studying computer science. He has never run the operating system he is criticising, nor has he even made the effort to obtain a retail copy of that software so that he might validate his journalistic credibility.
Please people, I realize this article is somewhat more rambling and pointless than most, but isn't it a little inappropriate to give a public forum to someone who self-confessedly doesn't know what he's talking about?
-konstant
-konstant
Yes! We are all individuals! I'm not!
I think the article makes some very interesting points. However, one thing I have found that is uniform with people who claim success with NT is an aggressively restrictive attitude towards what users are allowed to do.
Arguably, this must be true of any successful widely deployed solution. Since most people don't have the disposition or training to be an effective system admin, then they should be locked out of certain administrative functions. On the other hand, the standard policies of IT departments seems to be almost draconically restrictive -- for example no user installed applications (of course these days they should probably count any file with a ".doc" extension as as a user installed application).
This seems almost like a 19th century attitude towards workers -- like central planning or Taylor's discredited "Scientific Management" theories. Speaking as a former MIS director, I feel that users should on one hand not be required or in many cases allowed to perform many administrative functions, but on the other hand they should be free to customize their toolsets. It makes no sense to call somebody as "knowledge worker" and then dictate how they will do their work. A modicum of chaos is a healthy aspect of any complex adaptive system (e.g. an enterprise).
An ideal OS would provide a customizable mix of control and freedom. But both the control and freedom need to be targeted at freeing users to be as effective as possible. Maybe its just my personal disposition, but users of a system should feel free rather than restricted. What I question is whether an NT deployment can be both manageable and provide this sense of user empowerment.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
Unfortunately, it is exceedingly difficult to find any unbiased voices at /. anymore. With the whole WinNT vs Linux thing, all the posters here have presented anecdotal evidence as to why NT is so much worse than Linux. Very few people come to NT's defense, not because no one suports NT (they sell enough licenses) but because any NT users who might have been here have left.
You see, filled only with pro-Linux zealots, /. becomes little more than a group of the same people patting each other on the back. Very little is accomplished in saying "NT sucks, here's why" without having anyone here to defend it. Yet time after time, when someone jumps to NT - or any non-open source competing product for that matter - invariably they are pounced upon by fellow /.'ers.
It feels like its not really worth airing opposing opinions on this forum anymore. In this mob rule, one's dissenting voice gets lost in the incessant bickering that accomplishes very little. For those of you who have said that NT or whatever is difficult to use, ask yourself ... no, really ask yourself ... where the computer industry is heading and why.
I work for a small state agency where we have three people working in IT. Two people handle the network, servers, and workstations, and I handle the web servers. I manage content, program, handle security, and do a myriad of other tasks to keep my four servers up and running. Two of the web servers run WinNT, the third runs Luinx, as does the Database server. "But why would you run NT, when Linux is clearly superior?" I'll tell you. The linux box has been in this agency since about 95 or so. Its running whatever version of Slackware was out at that time. Its been patched along the way (I hope...), but was the baby of an employee who no longer works here. He didn't leave documentation. One of the two NT servers was also here when I got here and was also the toy of another employee. Which one was easier to take over? I hate to say it, but the NT server was, because it was easier to assess what was installed and what needed to be done with it. The Linux box is sitting there, waiting until I have a solid week or so to go through and see what needs to be done.
NT addrssses a very important issue with the people at my job. Because the IT staff is very small, and the network admin is a NetWare freak while the desktop person knows hardware much better than software, the systems have to easily transfer from one person to the next. I'd be doing a disservice in moving from an NT webserver because the next person they hire will need to quickly assess what they have/what they have to do. Or, because of budget cuts, they don't hire someone, they're going to have to train someone on the basics of how to handle a web server. I can't do that with a Linux box as well as with an NT solution.
Then why a Linux DB? First, it was cheap. Second, the interface for the DB resides on the Linux box. Third, I reason that if the person might as well learn simple linux stuff if they're going to administer a database.
Most of the Linux users out there would rather die than use a GUI to get a job done. They'd much rather perform hand calculations and do everything with a small set of utilities. Whether or not the level of graphical automation has anything to do with the quality of the output is up to speculation. Despite the existance of the Linux command line for the last 9 years industries which only existed in the last 2 years have defaulted to consistantly producing TV shows on graphical interfaces running Win NT.
Either that or they are blind to deficiencies in their setup that will only appear as the setup gets hammered a bit more.
A problem I'm seeing more and more of as linux and nt become more widely used in our environment is that since most users' first experience with both of these OS's was on a home machine, they become extremely lazy in their sysadmin practices.
A sysadmin or developer who may be fairly clueful and careful normally, suddenly becomes a install-happy fool on a redhat or nt box. While previously they would have no problem spending a couple hours reading docs on new package, now suddenly they have no problem just running through the install wizard or installing the rpm, and calling it done.
While this isn't *too* bad on the linux side, because in the back of their mind they know they're being lazy, its devasting on the nt side because for many users, this is the only way of administering an NT box they've ever known. And since it seems to work for them at first only using the wizards, they keep doing it. And they get more comfortable with it, and keep doing it... even if it does catch up with them, and they trash the machine out of ignorance, they just chalk it up to NT's instability, because that's all they read about on slashdot. =)
This sounds ludicrous, perhaps, but how many production NT boxes out there do you still see %systemroot% still with everyone:f? Or how many nt admins give you a blank look when you ask what options permissions they assigned to that new service they just installed? Quite a few, I'm afraid.
NT definitely needs time to mature -- the fact the ms threw away nearly *all* the nt4 mgmt tools shows that even they realize this. But just as important, the sysadmin community needs time to mature with respect to NT and grow more clueful admins. Interns-turned-nt-admins, or unix guys who have a quake machine at home and bought a learn-nt-in-7-minutes book will probably be able to get an nt box installed and most of the services a company may need up and running, but when things go wrong, about all they'll be good for is to fetch the os media and reinstall, complaining about microsoft's evilness the whole time.
There is no replacement for a clueful sysadmin, on any platform. No matter what the little bullet points on the vendors web site say. The fact that NT has alot of vms-ness underneath really should drive this home -- if an admin can't tell an sid from a guid, or gives a blank look when you mention \??, then it's only a matter of time before they get knocked on their butt by this "simple-to-use" os.
Clueful admins aren't grown overnight, and they they don't pop straight out of msce's courses or trade schools, so I don't mean to criticize anyone who's struggling on this platform. But it's very easy platform to gain a false sense of confidence in your knowledge of it.
In my opinion, this is it's most fatal flaw, and this is an excellent article to point it out. Hopefully with microsoft improving the docs and adding obviosuly complex things like active directory to NT more admins will realize how deep they are in over their head and seek out deeper understanding of whats going.
Anyways, that's my rant for the morning. Time to go home and sleep.
-- Scott
I usually stay away from the *nix & holy war but this article is of particular interest due to my observations around the project I am working on. I've been builing a test lab for a large northeast company that offers both the Solaris & NT platforms for web hosting. This lab was to be a replica of the production environment down to the routers / switches used. In building this lab, I had to work with the Unix & NT Web Hosting engineers. During this time, I received awesome exposure to Solaris and Linux (which I was using for the development platform). My learning curve was vertical (I come from a NetWare and NT background). Needless to say, I was stoked at the learning opportunity. Now to my point. Initially, the Solaris builds were tough. I'm a CLI and script guy so the UI wasn't the problem. It was the lack of readily availabe and accurate documentation that was painfull. I can't count the number of contradictory HOWTO's and FAQs that I ran into. However, after working with some of the Unix sysadmins who "knew" their stuff, I "knew" what I needed to look for. I scripted the builds and everything took off from there. When I worked with the NT engineers, it was equally frustrating. The NT build was done through a series of WinInstall packages and PERL scripts. I had to modify the NT build in the lab due to trust dependant security. When I asked for the dependancies, most of the engineers (MCSEs)didn't know what the scripts they were running did! I finally found someone who "knew" (non MCSE)the plumbing of the builds and I was off and running. A platforms ease of use is relative to how much the sysadmin "knows" about it. From my limited experience, *nix requires you to get to "know" it before it is up and running (this is a good thing!). Unfortunatelty, NT allows you to get it up and running without you "knowing" what is going on (NOT a good thing!). If the NT sysadmin "knows" his stuff, it can be a decent and stable platform - as can NetWare and the *nixes etc ... Any worthy sysadmin should know that easy isn't always better - although you shouldn't have to spit blood for everything either!
"Hatred is the coward's revenge for being intimidated"