Looking for a tool to make plain text document out of a simple HTML document

[Akseli Mäki]

Hi,

Hopefully this is not too much offtopic.

I'm working on a FAQ. I want to make two versions of it, plain text and
HTML. I'm looking for a tool that will make a plain text doc out of the
HTML doc. The HTML version doesn't have anything fancy, just internal
links. So the tool must be able to delete internal links and anchors from
the HTML version, but leave external links in simplified form. That is, the
HTML version would say <a href="http://foo/bar.html">Bar</a> and the plain
text version would just say http://foo/bar.html. The tool should also be
able to make mailto links into plain text, that is, change <a
href="mailto:fo *@bar.com">Foo </a> into fo*@bar.com. In fact, I think both
of these changes might be possible with regex search and replace engine. I
have one at my editor, but I don't know how to use it. So I'm either
looking for a regex strings or a ready tool, both will be find.

[Alan J. Flavell]

On Mon, 22 Dec 2003, Akseli Mäki wrote:

Ok thanks for all the suggestins. I already have Lynx so I'll use it.

Yes, Lynx does that job very well - except for tables. Real tables, I
mean - it can be actually beneficial what it does with
tables-for-layout, but tabular data can become unusable in Lynx.

(If you're in control of the HTML, I have an ancient web page
about how to make HTML tables which also present acceptably on
Lynx, but it's not really been updated since 1998: if you still
want to read it after that low-key introduction, then
http://ppewww.ph.gla.ac.uk/~flavell/www/tablejob.html

The nobreak-space stuffing technique is the one I would recommend now,
if you want to do anything at all.)

[David Dorward]

Alan J. Flavell wrote:

On Mon, 22 Dec 2003, Akseli Mäki wrote:

Ok thanks for all the suggestins. I already have Lynx so I'll use it.

Yes, Lynx does that job very well - except for tables. Real tables, I
mean - it can be actually beneficial what it does with
tables-for-layout, but tabular data can become unusable in Lynx.

Lynx actually does a pretty good job with tables these days, it guesses if
its a layout table or a real table, so its not 100% though.

To take the example from the URI I snipped, it comes out quite happily as:

Deutsch British USA
Haube Bonnet Hood
Kofferraum Boot Trunk
Benzin Petrol Gas(oline)

(With the <th>s rendered in brown)

(That said, it is rather early here, and I only skimmed your document, so I
could be missing something).
--
David Dorward <http://dorward.me.uk/>

[Nick Kew]

In article <Pi************ *************** ****@ppepc56.ph .gla.ac.uk>, one of infinite monkeys
at the keyboard of "Alan J. Flavell" <fl*****@ph.gla .ac.uk> wrote:

Yes, Lynx does that job very well - except for tables. Real tables, I
mean - it can be actually beneficial what it does with
tables-for-layout, but tabular data can become unusable in Lynx.

Um - recent Lynx does a very nice job on tables.
At least, that's my experience with the Lynx bundled in Slackware 9.

--
Nick Kew

In urgent need of paying work - see http://www.webthing.com/~nick/cv.html

[Alan J. Flavell]

On Mon, 22 Dec 2003, David Dorward wrote:

Lynx actually does a pretty good job with tables these days, it guesses if
its a layout table or a real table, so its not 100% though.
OK, I knew they had been working on it. Care to mention which version
you're using?
(That said, it is rather early here, and I only skimmed your document, so I
could be missing something).

OK, I *did* stress that my page is from about 5 years back. However,
I might remark that there used to be plenty of obsolete versions of
Lynx around. Hmmm, interesting: if I trawl the logs now, the oldest
version of Lynx that seems to be well represented is 2.8.3rel.1,
although I did see just the occasional 2.7.1

A quick hunt around different machines here didn't show up too many
usable versions, but I tried a couple as below.

This one doesn't space the display out usefully: Lynx 2.8.2rel.1

This one does: Lynx 2.8.4rel.1

So it must have got sorted out somewhere in between.

I think the bottom line is that I withdraw my previous posting,
provided that the user has a recent-enough version of Lynx. Thanks.

[David Dorward]

Alan J. Flavell wrote:

On Mon, 22 Dec 2003, David Dorward wrote:

Lynx actually does a pretty good job with tables these days, it guesses
if its a layout table or a real table, so its not 100% though.
OK, I knew they had been working on it. Care to mention which version
you're using?

david $ lynx --version
Lynx Version 2.8.4rel.1 (17 Jul 2001)
libwww-FM 2.14, SSL-MM 1.4.1, OpenSSL 0.9.7c
Built on linux-gnu Dec 7 2003 10:16:09
I think the bottom line is that I withdraw my previous posting,
provided that the user has a recent-enough version of Lynx. Thanks.

Its always nice to see browsers improve. (Hint to Microsoft :D)

--
David Dorward <http://dorward.me.uk/>