[sw-l] Frustrations of SignWriter DOS users...

[Sandy Fleming]

Hi Val!

> Will there be a day, when Stefan will be able to type full documents
> directly in the visual symbols of SignWriting, like we do in SignWriter
> DOS? We can type directly without ever seeing one word or letter in a
> spoken language...pure SignWriting documents that can go for pages and
> pages...that kind of typing, in a sign-processor program, that can also
> save signs into a dictionary and paste signs from the dictionary while
> typing...and then be able to print both documents and dictionaries in
> multiple print formats...those features have never been done in another
> computer program...SignBank gives us some of that, but does not have
> the typing element...

I thought I'd try to explain a few things to put this in perspective.

The real pity about SW-DOS is not that it's dying but that it lived so long.
It would have been better if SW-Java had worked out and we had had a modern
version of it. No matter how much anyone may have come to love SW-DOS, or
how useful people have found it, there is probably no longer any practical
way to expand it beyond the 10,000 limit, or make the screen as attractive
as a modern word processor (say, by offering better layout, colours and so
on), or achieve any sort of advancement at all. It's not just Bill Gates
doing things, it's that modern software development processes really are
better and programmers quite rightly think twice about trying to bring a DOS
program up to date - there comes a point where it starts to get easier to
write a new thing from scratch than to try to keep improving the old one.

The question was asked of whether the current suite of Web-based programs
would ever be used to create large documetns, and the answer is no, not any
more than anyone would try to use a browser connection to implement a
powerful word processor. Browser connections are slow and really have very
litle - almost nothing - in the way of memory. This lack of memory is a
deliberate feature of HTTP design, and so the answer is no, the Web is not
really suitable for creating large offline documents.

So you might ask why Stephen uses Web-based development for these things.
Part of the answer is that his aim is mainly to enable everyone all over the
world to build dictionaries (not write documants) and for this a Web server
with a database is ideal. Another part of the answer is that browsers and
servers supply a lot of programming power and very good database and user
interfacing software just by default. This makes it possible for Stephen to
respond to requests for improvement very quickly: he can add new features
without having to ask everyone to download and reinstall anything. A
programmer writing a word processor for offline use would have to wait and
save all his changes in a new issue, which he would have to test very
heavily to make sure there were no errors before asking everyone to
reinstall the new version. Web development, however, enables Stephen to
interact with us an respond to our ideas quickly. He can experiment without
having to worry about wasting peoples time with downloads and installations.
In fact I think he added a new search feature this morning while I was
working on the BSL dictionary, and I didn't even have to stop work for a
moment, never mind download and reinstall everything again!

This is all very valuable because as we've seen Stephen has been developing
at the very heart of SW technology, trying out our ideas, and these
developments can be used in future word processors when programmers do
develop them. The SW word processors of the future are bound to be much
better than they otherwise would have been, due to the sort of work Stephen
is doing now.

I would also emphasise the importance of developing good XML (ie SWML)
applications. XML is not only very easy for programs to read and write, but
is also very easy to transform (or operate on). This means that it doesn't
matter that we have different dialects of SWML - it's easy to transform one
dialect into another. It's easy to upgrade because if we add new features or
remove old ones, we can write an XSLT stylesheet to transform all out old
data into the new form. This means our data will last forever even if we
keep wanting to keep improving on the way it's stored, unlike the old .sgn
data which was a fixed format and difficult to change without some serious
programming effort. If this is too technical, I could just wrap it up by
saying that the whole idea of using XML (ie SWML) is to make sure that once
a sign is in a dictionary, or once a document is written, it can be used
again and again and changed again and again by as many different
applications as programmers care to write.

So to summarise, the web-based development currently taking place is a very
VERY good thing indeed, but don't expect it to end up by giving you a
full-featured word processor. That's a separate programming concern and
involves much longer-term development. But the new breed of SW word
processors will be all the better for their workings and data having been
poineered in these Web applications.

Phew! And I thought I'd get an early night tonight!  :)

Sandy