http://www.w3.org/WAI/ER/existingtools.html

I've found and read   http://www.w3.org/WAI/ER/existingtools.html  - the
page looks very useful.

(1) How about a "maintainer" E-address for matters not necessarily
needing the attention of a whole mailing list?

(2) IMHO, W3 being international, dates would be better in international
(ISO-8601) format : not 2/25/2000, but 2000-02-25.  The first form is
ambiguous for 11 days of the month.

(3) Currently, the cited Pehtoori service is not responding.

(4) I use W3's Tidy as a checker only; your page does not show that
possibility.



(5) The main point :-

I have written, for my own benefit, and made available, a program to
check local links within a [master] copy of a Web site.  That is, it
scans a given HTML file, and for each "<a href ...>" without a protocol
indication (*:) it tests the existence of the file and anchor; and it
recursively scans those files (if the extension is right) ... .

It is written in Borland's Turbo Pascal 7, so is for 16-bit DOS, but it
also runs in Delphi 3 (32-bit) console mode.  It might, therefore,
become usable with "Kylix" for UNIX.  It compiles with (free) TMT Pascal
demo v3.50, but then runs unsuccessfully, reason unknown.

It is intended for use on pages which are already valid HTML; it does
not parse the HTML properly, but scans for what it wants with apparent
success.  A proper parser, to locate the relevant constructs, would be
more elegant.

It is fast: a few seconds on a PII/300 to check my own site master; the
output ends :

Summary: 66 files tested ; 535 anchors seen ; 2131 relative cites seen ;
 Relative citations:  directories - missing 0, found 7 ;
 Relative citations:    filenames - missing 0, found 285 ;
 Relative citations:      anchors - missing 0, found 1029 ;
 Local URLs 12 ;  Odd A-refs 0 ;  Make relatives 0 ;  Unused NAMEs ? ;
 Repeated NAMEs 0 ;  "\" in links 0 ;  Links exceeding 8.3 format 0 ;
 UnQuoted HREF/NAME/&c. 1 ;
 Far URLs 1563 ;

There does appear to be a counting error, since "cites seen" is greater
than the total found + missing.

I prefer to use it in a DOS window, with pause-on-error, fixing errors
with a Windows editor as they are found; link errors are usually minor
blunders, easy to fix on sight by the blunderer.  But it uses standard
output ...

At present, it recurses enthusiastically down every file it sees, eating
stack; the 16-bit version handles my site when starting at the core, but
when starting at the fringe it needs more stack (32-bit: OK).  I am
slowly working on a modification which, by using more lists, will
iterate more economically.

The program is uploaded, after each useful change, to my Web directory
<URL: http://www.merlyn.demon.co.uk/programs/> as cheklinx.pas &
cheklinx.exe; also both together as cheklinx.zip.  The only instructions
are in leading comment in the pas file; I might do an associated txt
file.

Dave Raggett has heard of it.

There is a list of "questions" on your page :
Name                                           CHEKLINX
Developer                                      Dr J R Stockton
Description                                    Relative links checker
Does it run locally or on a server or both?    Locally
What platforms are supported?                  DOS (and Windows DOS box)
If server-based, is the server software available for installation on
    local intranets?                           N/A
Is support available to users of the software? In moderation only
Was the software designed to be accessible to all users?
                       I see no problems; but "designed" hardly applies.
Date of last update, if available.             2000-09-08 release.

 
-- 
© John Stockton, Surrey, UK.  jrs@merlyn.demon.co.uk   Turnpike v4.00   MIME. ©
 Web <URL: http://www.merlyn.demon.co.uk/> - FAQqish topics, acronyms & links.
 PAS, EXE in <URL: http://www.merlyn.demon.co.uk/programs/> - see 00index.txt.

Received on Monday, 2 October 2000 14:43:35 UTC