Hi Lex,

> The last time I played with RTF was a few years ago, but it was a bad
> experience.  I went on a jihad asking everynoe to use RTF, because it
> seemed like such a nice solution to the problem of portable word
> processing formats.  However, it didn't work out.  People sent me RTF's
> that none of my tools could could read.  The problem was that RTF files
> generated by MS tools tended to included MS-specific objects in them.
> You'd be going through the file and see messages along the lines of
> "unparsable object  excel210413".

According to the specification you can ignore all the tags you can´t manage.
You can handle them as binary contents (as an image) and save them
unchanged.
This kind of approach let you handle the elements you understand and also
save multiple copies of the same data (in more than one representation, as
used by MSWord with embedded images).
If you ignore unknown data you will not loose anything (they are not objects
:-) ).

This kind of flexibility can let us store proyects and any squeka object in
RTF tags and handle/share active contents in RTF.

> Has this situation really changed?  If not, then we can never expect to
> reliably import RTF.  We can export stuff that is strictly conformant,
> but  there are many of the existing tools that export stuff we will not
> be able to read.

This has not changed in the last 10 years or so.
I have implemented some solutions for bussiness using data exchange in RTF
format (for serving "objects" to MsWord, used as rich text formater for
corporate information) and they has been working without mantainment in the
frameworks that handle RTF.

> To contrast, there are plenty of open formats around, many of which are
> older and better established than RTF.  For word processing, there is
> TeX, SGML, HTML, and gee, plain old TEXT.  For output, there is PDF,
> PNG, and JPEG.  For Squeak-to-Squeak interchange, you can't beat project
> files.  These are all formats that we can definitely read.

The current implementation of PDF is only a draft version and does not
handle most of the PDF specificacion.
The PDF format is more complex than RTF specification (10x1 ratio in
complexity... I have implemented frameworks for reading PDF and writing RTF
and I think that the complexity of PDF spec is important).

One of the weakness of RTF format is lack of compression.
Documents in RTF consumes a great amout of space in disk and in memory.
PDF is better for document compression and it is used for documents with
high resolution.

> Further, there are plenty of non-technical reasons to avoid things under
> MS's sponsorship and control.  At any time MS may pull the plug and
> start emiting non-standard content, and they already have a long history
> of doing it!  MS does not play nice, and we should stay as much out of
> their power as possible.  Imagine if we do try to standardize on RTF,
> and MS yet again turns the tables and emits something we cannot parse.
> Imagine lobbying Microsoft to get something fixed, and being powerless
> to fix it ourselves.  Yuck!!

We can imagine the same and change the name M$ for any other...
(nothing is so bad that it can´t be worse)

cheers,
Ale.