On Tue, Jan 28, 2003 at 10:22:25AM +0100, Tim van Erven wrote: > On Mon, 27/01/2003 09:25 -0500, Daniel Barclay wrote: > >> From: Edward Craig <[EMAIL PROTECTED]> > >> Well, some of us still run to terminal mode, but it's all based on > >> Hollderith cards, I tell you, all because of IBM and the IRS. > > > > Well, there's also a little factor from how many characters in a readable > > font size fit across a standard size sheet of paper. > > Ever noticed how many characters there are on a line of a newspaper or > in books? It may nog be exactly 80, but it's close. The reason is much > longer lines are harder to read. Try putting some tekst on a page in > landscape; it's really annoying.
Actually, it's closer to 60 characters. That's why LaTeX wraps at about 60 characters by default. Typesetters decided a long time ago that lines shouldn't be longer than that. Newspapers are even smaller (about twenty). They're typeset that way for a reason; such short lines are very easy to read quickly because your eyes don't have to re-focus multiple times per line. Having a limit to the number of characters per line is very important, unfortunately 72 is a bit too wide. That's because, of course, the standards were written by technical people, not by typesetters. They set the limit at that width because it would fill their 80 column terminals nicely and still allow for some quoting, rather than at the widest width that's easy to read. The terminals were made wider than would be the most optimum for prose, I imagine, because 60 characters really isn't wide enough for code or formatted data. I think 8.5in wide paper being as wide as it is is a horrible shame as <60 character wide text at a reasonable font size has ridiculously wide margins. I have noticed that court opinions are written single column and about 60 characters per line, and many textbooks have wide pages with very wide outside margins in which they put text and notes, but people generally just end up stretching the lines out so the page isn't half-empty. I suspect that the paper is so wide because it was designed for hand-written messages. When typewriters were invented, allowing the masses to write typewritten text, everyone just kept using the same paper on which they used to write by hand. On the subject of wrapping lines, of course modern mail readers can wrap long lines. Hell, my TERMINAL can wrap long lines so I don't lose data off the edge, but that still means things end up looking like crap when it finally reaches the newline and it's not aligned with edge of the terminal. I think someone suggested that the author doesn't include any newlines in the paragraph so the reader's MUA can wrap lines where he specifies. There are two problems there. First is the previously mentioned problem with quoting. I've noticed that MUAs that don't insert newlines at the end of their lines are the same ones that promote horrible quoting practices (ie, composing the entire previous message after a -- Original Message -- line). Honestly, at that point the quoted text has become so useless that it might as well have just been excluded from the reply to save bandwidth and storage space. The other big problem with the no-newlines idea is that the sender loses the ability to apply special formatting to her message to make ASCII art or tables (for instance). She has no idea where the reader's MUA is going to wrap the lines so her neat diagrams and tables could end up being a bunch of incomprehensible gibberish to the reader. You therefore end up *losing* functionality without significant improvement by eliminating the end-of-line newlines. Another argument I've seen is that few use people use 80 column terminals anymore so it's stupid to have such a restriction and end up having all of that wasted space to the right of the text. Granted, most people have the *ability* to look at longer lines, but that doesn't mean they should. I prefer to use that extra space on my screen to tile terminals so I can look at multiple terminals at a time. I'm currently typing this in a 80x59 emacs window with two 80x24 line aterms to the right of it. Besides that, if not at 80 characters, where *should* it wrap? If anything, the previous discussion of line widths and readability suggests that lines should /shorter/, not longer. The best reason I can think of for wrapping at <=72 characters is that the standard requires it. The *only* appropriate place for this discussion would be in an ITEF working group for the purpose of updating the standard. Until RFC 278 (and all other RFCs that mention the 72 character limit) are obsoleted by a standard that suggests something different, or until we quit using plaintext messaging, then MUAs must wrap the text they send at 72 characters. That's all ye know on earth and all ye need know. -- Jordan Bettis <http://www.hafd.org/~jordanb> All things that are, [President Jimmy] Carter told Americans the truth and they hated him for it. [They] responded by throwing him out of office and replaced him with a movie actor who promised to restore the Great Enterprise to all its former glory, whatever the costs. -- James Howard Kunstler, _The Geography of Nowhere_
