Jack Campin escreveu:
>
> > 5. Line terminators
[...]
> BBEdit Lite does the same (and the full version of BBEdit also manages
> site uploading for you). But the problem isn't with text processors,
> it's with brainless transfer agents that don't convert terminators in
> text files properly. If the original file uses any of the standard
> options coherently, it's not the source site's problem if duff web
> browsers and mail agents fuck up.
>
John Walsh escreveu
> I have a bunch of files that I've transferred between Unix and Dos,
> and edited on both, so they have some line endings from each. (And
> it'll get worse if I ever get that Mac notbook...)
In fact the flaws are all along the line, IMHO:
1 ABC relies on line terminators as a meaningful unit of coding. SGML
(and its offspring HTML), on the contrary, avoids relying on them
exactly for reasons of portability, but even so has an algorithm to
handle them in the couple of circumstances they might be a point (please
see Martin Bryan's "SGML and HTML Explained" chapter 11, section "11.4 -
The effect of record boundaries" at <url:
http://www.personal.u-net.com/~sgml/book/home.htm#Contents>).
2 Text editors should allow you to keep consistency in line terminator
formatting upon request (or not, if for some reason the file format you
are typing assigns different roles to CR and LF and their combinations),
instead of assuming just one of several possibilities. Some try to
guess the format regardless of the local system's default.
3 ABC software don't seem to be generally aware of platform
specificities. They should take portability issues into account and
process line terminator candidates one by one trying to make sense out
of them to allow for mixed platform editing � la John Walsh and minimize
the risk of misinterpretation of the author's intention. E.g.,
<other><cr><lf><cr><lf><other> and <other><cr><lf><lf><other> would be
both interpreted as <other><line end><line end><other>, but there would
be no solution for cases in which a <other><cr><other> further received
a <lf> upon editing on another platform (resulting in
<other><cr><lf><other>). Again, that is the consequence of not having
an explicit code to signal meaningful line ending in ABC language.
I think it's too late to change the way ABC encodes functional line
termination; also we cannot expect the other segments (text processors,
network transfer software) to deal with ABC specific issues. So, it's
up to ABC software developers, who are concerned with these problems, to
manage workarounds where due.
Regards,
Paulo Eleut�rio Tib�rcio
To subscribe/unsubscribe, point your browser to: http://www.tullochgorm.com/lists.html
