Cecile Chi sent me this message:

> I'm using QCTerm .70, which was supposed to have a fix for a VPlus
>  problem in .65.  I don't know what the problem was with .65, but I have
>  a problem with .70.  I ran FORMSPEC and made a slight change to
>  one form.  I shifted a field by deleting a blank several spaces after the
>  field and inserting a blank several spaces before the field.  This screen
>  has 88 fields on it, all delimited with hidden brackets.  I used ENTER
>  to save my change and then recompiled the forms file.  No errors.
>  When we tried to use the application program that uses that forms
>  file, everything was fine until we got to the screen I had changed.
>  All of the hidden brackets were gone, so the screen appeared just
>  as it was designed, except there were no unprotected fields.
>  To fix it, I had to use Reflection to run FORMSPEC and then put in
>  all 88 beginning brackets and all 88 ending brackets and then redo
>  all the highlighting and field definition instructions.  Perhaps you
>  should warn your users not to run FORMSPEC through QCTerm.

Art Bahrs sent a similar message a few days ago. While we don't begin to have
enough information yet to understand what the problem is, it is probably wise
to use one of the other terminal emulators for FORMSPEC design at the moment.

We're still finding a few very small bugs in QCTerm. However, for the great
majority of uses, it now seems to be stable enough to be distributed to
lightweight users. A normal user will probably never see a problem.

The errors we're getting now are all application specific, simply because the
failing programs did something outside of the domains that we've thoroughly
tested and are hitting something that we either screwed up or misinterpreted.
Speedware screens that make use of multiple TABs to locate a drawing position
on a screen rather than escape sequence addressing are our currently most
outstanding problem. We know what the problem for these screens is and we
will correct it as soon as possible.

At the moment, we do not understand, nor can we even repeat, the problem with
FORMSPEC, but hopefully we can find a test case soon that will allow us to
repeat the error so that we can diagnose it and correct it as soon as
possible.

In the interim, if you're going to do FORMSPEC design, it's probably wise to
use one of the older emulators. Your users should be fine with using QCTerm
for the resulting forms. We have no indication that the end-user forms are
failing for anyone.

Wirt Atmar