Re: gEDA-user: PCB: Stale rat's nest?

[sdb@xxxxxxxxxx (Stuart Brorson)]

> Since you guys asked about bugs, here are some things I ran into=20
> while using gEDA and gSchem.. =A0Some of these aren't necessarily critical =
> bugs,=20
> but they all affect the useability of the program, some quite severely=20
> compared to commercial offerings. =A0Here goes, very long list follows:

Thanks for taking the time to do this.  The list is very interesting.
Personally, I would divide the bugs you list into 4 catagories:

1.  Real bugs.  For example, I agree that gschem should not start up
zoomed waaaaay out.  This has been noted before.   The fix is easy, I
think.  Also, making gnetlist recognize --help is a good catch, and
should be easy. 

2.  Non-bugs having to do with the fact that you didn't read any
documentation first.  I wonder if we could somehow make Bill Wilson's
excellent tutorial pop up the first time somebody ever tries to run
gschem?  I've been meaning to make it pop up at the end of a CD
installation. . . . 

Note that I do belive that you *shouldn't* need to read documentation
to make a program work.  I never read documentation, for example,
unless I am stuck.  Therefore, I do think that we could make the tools
a little more self-explanatory.  However, you appeared to be stuck
more than once; a quick Google search might have gotten you out of
your corner.   

3.  Bugs in "geda", the deprecated project manager.  The wiki,
discourages its use.  We either need to fix it, or remove it from the
distribution.  Had you read the documentation first, you would have
found that out.   But, yes, you shouldn't need to read the docs to
discover this in the first place.

4.  Issues having to do with the philosophy of gschem's (or any EDA
tool's) usage.  You may not know that different tools use
different philosophies.  Here's one which I don't think gEDA should
change:  

> * gSchem should never require the user to manually give each part a unique=
> =20
> name, e.g. U1, C5, R22... =A0While it is somewhat easy to renumber all part=
> s of=20
> a specific type, it should not be necessary for the beginner to use that=20
> function.

Orcad, for example, also does auto-refdesing.  Viewdraw, however,
does not.  I personally want to set my own refdeses, not the program.
Therefore, this is a feature, not a bug.  This behavior shouldn't
change. 

There are two refdesing utilities available: refdes_renum and grenum.
Also, you can configurably have auto-refdesing performed within
gschem by setting up your gschemrc (or system-gschemrc) file
correctly.  Perhaps we developers should discuss making autorefdesing
an automatically installed feature which power users can disable.  Or
maybe make it configurable from within gschem via a pull-down, but
set it disabled upon installation.

(Indeed, this brings up the whole interesting idea of making the
various settings in gschemrc settable within gschem via a pull-down
menu item called "settings".   Hmmmm . . . . .)

Stuart