Re: website & identifiers

[Thomas Dickey <dickey at his dot com>]

On Sat, 27 Dec 2003, Harold L Hunt II wrote:

> Thomas Dickey wrote:
>
> > When perusing the website for news, it would be useful if the pages were
> > marked (in their source) as generated or manually updated.  For the former
> > (unless they're generated on demand), a modification date would also be
> > useful.  For the latter, an RCS/CVS/etc identifier to distinguish
> > successive versions is normally expected.
> >
> > Are the pages all generated from another format, or are some in CVS
> > (somewhere)?
>
> By the way... what are you looking for?  CVS identifiers wouldn't

If it's maintained (i.e., if the file isn't edited w/o checking it in),
it's a quick way to check if the file's been changed recently.  The
identifier also provides a point of reference to check if there are
changes that aren't committed.  Hand-coded dates are more readable (and
of course I use those where readability is a factor), but the automatic
ones are preferable for identifying distinct versions.

> necessarily help to determine that new information has been posted since
> commits for spelling fixes, grammar changes, dead link correction, etc.
> would cause irrelevant noise in the modified date tags.  That's really
> why I chose to put a hand-coded date at the top of the pages; that way
> people are notified when content has been modified in a meaningful way,
> but there aren't false-positives when minor changes have been made.
>
> Would CVS identifiers still be useful for whatever you are trying to do?
>     Maybe I'll start adding them as comments for starters but still keep
> the hand-modified date for each page.
>
> Harold
>
>

-- 
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net