Re: Arlo, a little QA comment regarding your interview withlinux.com

[Alan Shutko <ats acm org>]

John Sullivan <sullivan eazel com> writes:

> On the other hand, in theory anyway, hidden files are hidden for a reason --
> they aren't intended for users to futz with.

Not always.  They may be hidden because you don't always want to see
them cluttering up your ~ file listing, but are definately intended to
be user-modified or user-viewed.  Many of my dotfiles come under that
heading:

-rwxr-xr-x    1 ats      ats           259 Oct 20 09:39 .Xclients
-rw-r--r--    1 ats      ats           211 Jul 23 13:01 .Xmodmap
-rw-r--r--    1 ats      ats          6337 Oct  5 23:55 .Xresources
-rw-r--r--    1 ats      ats           517 Sep 22 18:57 .aliases
-rw-r--r--    1 ats      ats           257 Jun  9 09:32 .authinfo
-rw-r--r--    1 ats      ats            24 Jun  9 09:32 .bash_logout
-rw-r--r--    1 ats      ats           185 Oct  8 10:36 .bashrc
-rw-r--r--    1 ats      ats            38 Sep  1 15:59 .cvsrc
-rw-r--r--    1 ats      ats          9046 Oct 23  1999 .default.css
-rw-r--r--    1 ats      ats         16138 Oct 25 10:50 .emacs
-rw-r--r--    1 ats      ats         10258 Oct 18 15:03 .emacs-custom
-rw-r--r--    1 ats      ats          2238 Oct 23 10:13 .emacs.bmk
drwxr-xr-x    2 ats      ats          4096 May 18 11:30 .env
-rw-------    1 ats      ats           500 Jul 21 21:58 .fetchmailrc
-rw-r--r--    1 ats      ats         14486 Jun  9 09:33 .fvwm2rc
-rw-r--r--    1 ats      ats         18148 Jun  9 09:33 .fvwmrc
-rw-r--r--    1 ats      ats          4240 Oct 21 19:13 .gnus
-rw-r--r--    1 ats      ats            20 Jun  9 09:33 .inputrc
-rw-r--r--    1 ats      ats           559 Oct 13 16:12 .kermrc
-rw-r--r--    1 ats      ats            37 Jul  9  1999 .less
-rw-r--r--    1 ats      ats            67 Jul  9  1999 .lesskey
-rw-r--r--    1 ats      ats           580 Sep 19 10:08 .mailcap
-rw-r--r--    1 ats      ats           825 Sep 19 10:08 .mime.types
-rw-------    1 ats      ats           245 Aug 14 11:43 .netrc
-rw-rw-r--    1 ats      ats          1351 Jun  9 09:33 .plan
-rw-------    1 ats      ats       2361550 Oct 25 14:13 .procmail.log
-rw-r--r--    1 ats      ats           708 Oct 16 20:53 .procmailrc
-rw-r--r--    1 ats      ats           642 Oct 13 15:22 .profile
-rw-r--r--    1 ats      ats            67 Jun  9 09:33 .project
-rw-rw-r--    1 ats      ats         24929 Jul 18 23:26 .rpmfind
-rw-rw-r--    1 ats      ats            92 Oct  7 19:20 .sawfishrc
-rw-r--r--    1 ats      ats          3394 Jul 10 17:26 .screenrc
drwx------    2 ats      ats          4096 Oct 17 23:51 .ssh
drwx------    3 ats      ats          4096 Jul 29 15:04 .ssh2
-rw-------    1 ats      ats          2093 Oct 25 11:10 .xsession-errors

Furthermore, this is a choice that is very dependant on what I want to
do.  Normally, I don't want to wade through 136 dot entries in my home
directory, but when I do, I want to do so for this instance, and I
don't want to have to change global settings to do it.

This is the same thing as certain settings in Netscape.  I may want to
browse with java and javascript off, but turn it on for this page.  Or
I may want to override the font size for a single page because of the
horrifying mess that web font use is.  (For this, Netscape does have
menu items, but I've never managed to make them work.)

> Also, every menu item adds a little bit of complexity to the
> menus. If the program has 200 menu items, it is much harder to deal
> with than a program with 50 menu items.

That's very true.  Which is why it's nice that with MS Office,
WordPerfect, and other older apps, you can change the menus to
incorporate commands that you use and need.  It's not something that
every user uses, but those who do (like my wife) use it a _lot_.  This
is an important feature, but not normally implemented because it's
harder for the programmer.  The gnome-libs should, imho, provide the
infrastructure to make this easier.  (Ditto keybindings and toolbar
customization.)

But what about the increased support costs?  Who cares?  You clearly
have to make a compromise between support costs and increasing
people's productivity.  After all, support costs would be less if we
didn't have preferences at all and instead hard-coded behavior.

-- 
Alan Shutko <ats acm org> - In a variety of flavors!
16 days, 21 hours, 23 minutes, 6 seconds till we run away.
Some people only open up to tell you that they're closed.