X-Authentication-Warning: delorie.com: mail set sender to geda-help-bounces using -f
X-Recipient: geda-help AT delorie DOT com
Message-ID: <534933A4.7000202@mochima.com>
Date: Sat, 12 Apr 2014 08:37:56 -0400
From: Carlos Moreno <moreno+geda-help AT mochima DOT com>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.4.0
MIME-Version: 1.0
To: geda-help AT delorie DOT com
Subject: Re: [geda-help] Having some trouble again!
References: <alpine DOT DEB DOT 2 DOT 00 DOT 1404050517330 DOT 8305 AT igor2priv> <533F89FE DOT 2060808 AT mochima DOT com> <alpine DOT DEB DOT 2 DOT 00 DOT 1404050746030 DOT 8305 AT igor2priv> <534000E2 DOT 5080703 AT mochima DOT com> <alpine DOT DEB DOT 2 DOT 00 DOT 1404060625040 DOT 8305 AT igor2priv> <5342AF6B DOT 6010405 AT mochima DOT com> <alpine DOT DEB DOT 2 DOT 00 DOT 1404080523041 DOT 8305 AT igor2priv> <53445BD8 DOT 4070505 AT mochima DOT com> <alpine DOT DEB DOT 2 DOT 00 DOT 1404090557260 DOT 8305 AT igor2priv> <53452A49 DOT 8000701 AT mochima DOT com> <20140409114728 DOT GA2172 AT localhost DOT localdomain> <5348800F DOT 3070500 AT mochima DOT com> <201404112358 DOT s3BNwoEw009040 AT envy DOT delorie DOT com> <5348973D DOT 9020603 AT mochima DOT com> <201404120135 DOT s3C1ZVgH012034 AT envy DOT delorie DOT com> <5348B7CE DOT 1070107 AT mochima DOT com> <201404120354 DOT s3C3sNtU018282 AT envy DOT delorie DOT com> <5348BE56 DOT 8000106 AT mochima DOT com> <201404120422 DOT s3C4MPbM020295 AT envy DOT delorie DOT com>
In-Reply-To: <201404120422.s3C4MPbM020295@envy.delorie.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Added-Header: 
Reply-To: geda-help AT delorie DOT com

On 14-04-12 12:22 AM, DJ Delorie wrote:
> Lower case letters at the end of refdes's *are* allowed.  They're just
> ignored.  We can't report an error for your case and support the other
> use case at the same time.
>
> The text in the message log showed it was looking for a device "Q".
> Since you don't have a device "Q" this should have been a clue that
> something was happening you didn't understand.
>
> The documentation for the netlist file format (section 8.5 of the PCB
> manual) states that trailing lower case letters are stripped.

I was sure (after you explained the issue, that is) that
something like this would be the case.  But still, I think
this still falls in the category of a "misfeature/bug", in
the sense that it's not user friendly to rely on a series
of "stated-somewhere-else" rules.  I mean, software
can (and in general does) check a huge list of predefined
rules;  a human brain is not that good at it.  An effective
computer interface brings the issues and the rules to
the attention of the user as they happen or as they
become relevant --- as opposed to state a series of
rules and conditions up front and then rely on the
user knowing and remembering them all.

Sure, an electronics CAD suite is not intended to have
the proverbial "dumb user", but still.

> Other than dropping support for the other use mode (which we've talked
> about before), there's little we can do since it's intentional.

Two things come to mind.

(1)  So, lowercase letters are ignored.  I still see a bug:
Why didn't the workflow chain work in my case?   It
seems to be that they were *inconsistently ignored*.
So, my symbol, that I thought was called Qmult, was
indeed called Q;  why didn't one of the tools find it?
That is:  why did one of the tools ignored the mult
and the other one(s) didn't?

(2)  Perhaps more importantly:  why didn't the logs
include a big warning message (I guess in the PCB
Designer Log window) clearly saying:  WARNING:
Refdes Qmult contains lowercase letters and was
converted to Q.  This may cause unexpected
results or trouble.  (or something along those lines).


Not being familiar how the internals work, I'm still
puzzled about the first paragraph in your comment;
you can't disallow lowercase letters because they're
necessary for the slotted components;  but then,
they are allowed, but they are ignored.  Does the
software ignore it in all cases?  I guess not;  for
the slotted components, taking them into account
seems to be necessary.  But then, if the software
can make the distinction, it should be able to
selectively allow them. No?

Bottom line:  as a user of the software, I don't find
much trouble in the specification that lowercase
letters in the refdefs I use are ignored.  But I do
find a bug/misfeature as per item (1) above, and
definitely I still think that the software could and
should guide the user better (as per item (2) above).


Again:  don't get me wrong --- I'm not trying to sound
like an ingrate!  :-)   I'm extra happy that you helped
me figure out what I was doing wrong, and I thank
you for that!   This suggestion goes with all the best
intention to help improve the software!

Cheers,
Carlos
--