I apologize to everyone (Dan, Ben, Matt) who posted in the discussion group and never got any reply from me; I moved all my mail accounts to a new server over the past two weeks, and it turns out that I never changed the address of the mail server in Frontier. So all the notifications that get sent out when there’s a new posting have been going into the bit bucket. All fixed now.

Dan brought me to an article I never read, but should have: Photography in the Age of Falsification. Anyone who works with me knows how much I dislike this, even when my own magazine does it. It’s just dishonest, no matter what the intention.

I decided to try out Secure Shell (SSH) today — I downloaded OpenSSH, and it compiled and installed without any problems at all, which is a great thing in Linux-world. One turn-off, though: the catty argument that the OpenSSH people have started with the person who registered openssh.org. They (the OpenSSH) people want everyone to have a kneejerk reaction against the guy as a domain squatter (they even tried to manipulate people with a Slashdot “advisory” full of hysterical security warnings), but as is always the case, the story is much more complicated than that. Actually, Alex de Joode (the openssh.org owner) has bent over backwards to try to resolve the whole non-issue, to no avail.

Holy cow, Jenni Ringley has some fans. The five picture sets that she sold on eBay yesterday brought in $3,010.23; her bed is currently at $2,660.

Yeah, I know that given that this site is called OS Opinion I shouldn’t expect any broad sweeping statements of inclusion, but I have little tolerance for people who believe that their way is the only way. Clearly, this guy hasn’t put nearly as much time into a GUI environment as he has his command-line one; if he had, then he’d realize how silly his configuration-of-a-web-server example is.

I pretty much agree with the sentiments expressed by Tom Watson in his latest column, Justice Department Saves the Internet, Film at 11:

The Microsoft suit is significant for one, and only one reason (no, Scott McNealy’s hypocrisy doesn’t count) — it has permanently created a Federal presence in the development of networked software in the United States. And that means, of course, lots of lawyers getting lots of hourly fees to litigate in an area they clearly don’t understand.


I don’t completely agree, Jason. For individual tasks, the GUI and the command line are really about the same in terms of complexity and effort. The GUI requires lots of clicks; the CLI requires typing out information.

But for any kind of repetitive task, you can’t beat the CLI combined with a scripting capability. You can set up mass migrations, grant or revoke all sorts of similar-but-not-identical rights, and set it all up to happen on a scheduled basis (say, midnight Saturday). GUIs have some neat-o features (ctrl-click multi-select, for example) that I’d hate to give up in the right circumstance. But give me Perl, SMS Installer/WISE, or Windows Scripting Host any day over sitting there in the User Manager (or MMC).

• Posted by: Dan Hartung on Apr 7, 2000, 1:10 PM

Hmmm… I didn’t get email notification of this posting. Strange.

Anyhoo, I agree that the GUI isn’t the most convenient thing for some tasks, but it assuredly cannot be beaten for others. I think that the arcane things that you have to remember for a lot of CLI tasks are its biggest problems, and the fact that I can just point and click for those seldom-done tasks is awesome.

So, for repetitive tasks, you’re right, the command line and some form of scripting or automation is a great thing. For something like configuring MIME types in my web server, a one-time thing, the GUI is great; I don’t have to remember which files or Registry keys it all goes in, and I don’t have to worry about file formats and parsability.

I think that my big point is that there is no one good way, despite what the author of that piece wanted you to believe.


• Posted by: Jason Levine on Apr 7, 2000, 6:11 PM

Oh, yes, I do actually agree. The nicest thing about a GUI is that it can organize your tasks and options visually so that you just have to choose the right tab or window.

I’m not a huge fan of the multi-row tab dialog so enamored by Microsoft programmers, but there is utility in having a single “options” window where you can configure your software. But the very utility of this works against it: you can’t put EVERYTHING in, and before you know it you’ve got a system policy editor and a user manager and a dozen other tools that are really related but not obviously organized. Windows got a little easier to deal with when they released TweakUI, but even that tool didn’t go far enough for a wonk like me. And seeing this from the coding side as well, adding functionality to a GUI is much more difficult because you do have to visually rearrange things if something gets added. A command-line switch doesn’t require the same level of attention.

I agree with you, there’s no one good way. Having a Mac with no command line drives me as nuts as DOS having a command-line and no GUI. ;-)

• Posted by: Dan Hartung on Apr 8, 2000, 4:40 PM
Please note that comments automatically close after 60 days; the comment spammers love to use the older, rarely-viewed pages to work their magic. If comments are closed and you want to let me know something, feel free to use the contact page!