On Mon, Sep 01, 2003 at 10:19:18PM -0400, Mark Roach wrote: > On Mon, 2003-09-01 at 13:26, Colin Watson wrote: > > As it happens, setting the tabstop option to anything other than 8 does > > irritate me when editing files containing tabs. I like to keep source > > code within 80 columns, and mismatched tabstop settings have a habit of > > screwing that up even if the file doesn't mix tabs and spaces. I've > > edited files that were perfectly readable on an 80-column display with > > tabstop=4, and were obviously intended that way, but that were > > unreadable with a default tabstop. In languages where mixing tabs and > > spaces is merely a cosmetic problem, the situation is worse yet. Leaving > > the tab character to represent its historical default of eight spaces > > simplifies everyone's life and means you can spend your life on more > > productive things than worrying about different kinds of whitespace. > > I can see what you mean. That is part of why I try never to use tabs at > all, especially in python code. I think both options have their > appropriate uses, and that for python code where it is uncertain what > tabs vs spaces will mean in terms of the code operation, spaces-only is > a reasonable way to go. > > Am I still way off the mark? Have I missed some other inherent evilness > in converting tabs to spaces? (Mind you, I am speaking only in reference > to python coding which is what I thought the conversation was about. if > not, my mistake)
Well, I don't use Python, so I'm probably guilty of being a little off-topic. There's one project I maintain that historically used tabs for indentation before I came to it, and I haven't changed that since it would result in unreasonable diffs in revision control and make other things hard to track down. I actually rather like it in some ways: it forces me to keep the level of indentation down, which in turn seems to result in more generally readable code. Nowadays, on average I tend to use expandtab for new code, but converting tabs to spaces is still an operation that needs to be handled carefully with respect to revision control, so I don't apply it across the board. Cheers, -- Colin Watson [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
