W3C home > Mailing lists > Public > html-tidy@w3.org > January to March 2000

Re: the -m switch (was: Bugs/suggestions)

From: Rick Jelliffe <ricko@gate.sinica.edu.tw>
Date: Fri, 18 Feb 2000 00:29:08 +0800 (CST)
To: html-tidy@w3.org
Message-ID: <Pine.GSO.4.21.0002172356460.26440-100000@gate>
Someone (Matthew Brealey?) wrote:

> I'm not sure whether or not this is a bug in Tidy.  Whether it is or
> isn't, my  mind boggles at the idea of somebody who's clearly no novice
> running a utility whose results are as radical as Tidy's to overwrite 
> a file without first making a backup.  

What I found was good practise was this:
	1) on all new files, run tidy on them twice
	2) check that the final output is acceptable still
	3) if not acceptable, correct the original, and repeat
	4) else accept the correct file

The twice-tidied files are fairly conservative. And well-formed
enough that I can simply add new material to pages and re-tidy the
updated pages blind (in fact, for a while I would just do it using
a process that ran at night). 

Running tidy once does not remove all ideosyncracies. So I think for
larger sites in which you will have to maintain and update pages
regularly, it is better to use tidy individually on each page to
*detect* any gross errors. After that, the pages are safe for 
the nice automated approach. 

Rick Jelliffe

(Of course, if the site is very large and consistent, it may be
better to not store data in HTML at all: store as XML and convert
out to HTML. In which case, tidy is useful to test the conversion
script or to allow fix up slack output.)
Received on Thursday, 17 February 2000 11:29:19 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:38:47 UTC