Re: gEDA-user: Some ideas on gEDA - please comment

[Ales Hvezda <ahvezda@seul.org>]

Hi,


>I am using gschem for about a week now since I just started a small
>electronics project. I really like it - it works as easily as the Orcad
>program that I used some years ago. But to the point now.
>

	Cool. :)


[snip]
> - I see no standard way of adding the author name and email to a 
>   just created symbol (I guess some specific attributes would do the
>   trick; possibly the auth= one mentioned below). Adding this info 
>   could help communication on using the symbol, finding out why it 
>   has / doesn't have power pins, etc.


	That's a good idea, author= and email= are good candidates
for this.  I've added these to my local copy of the master attribute
document.


>
> - When I upload a symbol at
>
>   http://www.geda.seul.org/symlibrary/upload.html
>
>   and later on find an error in it, what happens if I upload it again?


	The previous version is overwritten.  


>   In other words: are the uploaded symbols being version-controlled
>   via CVS or otherwise? (CVS would be preferable I think). A


	They are put into a temporary directory and then are placed
via a script into the CVS repository (after I review them).  The 
symbols are certainly not automatically placed into CVS, as I get
files submitted which are not symbols.   Apparently there are people
out there who think it's "funny" to submit garbage.  


>   version attribute could be of use.
>
>   The sym/titleblock/cvstitleblock-1.sym file holds the date=, rev=,
>   auth= and fname= attributes, using the corresponding CVS tags, but 
>   those attributes don't appear in the Master Attribute Document at

	
	That title block was meant for schematics which are placed
into CVS.  The auth= etc... were not meant as the author of the symbol,
but rather the author of the schematic.  It's a cool idea for both
uses.

>   
>   http://www.geda.seul.org/docs/20021103/attributes/index.html
>   
>   (so are these legal?).


	Yes for that particular symbol, but in general, author= and
email= are now the prefered attributes for new symbols.


>
> - A "live" symbol tree at the gEDA web site would be something
>   worthwhile to have I think. I.e. uploaded symbols being available 
>   to everyone instantly. If the upload code does a gsymcheck 
>   (it probably already does), this should prevent polution of such a 
>   tree. Having said this - gsymcheck (20021103 version) is more
>   liberal than the conventions given in the Symbol Creation Document at 
>   
>   http://www.geda.seul.org/docs/20021103/symbols/index.html
>
>   So maybe a more strict checking would be required before such a tree
>   could be provided at the gEDA site.


	Yeah, certainly a live symbol upload/download would be nice,
but in order to get there: 

	1) Yes, as you mentioned, gsymcheck would have to be made much
	   much much stricter.  Something I am willing to do.

	2) Somebody else has to write the web interface and scripts to
	   maintain/drive a live symbol upload/download.  Or an existing
	   web interface (with logins etc...) could be adapted)

        3) the interface has to be secure and safe, especially if it 
           runs gschem/gsymcheck.  

	I certainly do not have the time to write such a thing (would
be a fun project though), as I am already spending a huge chunk of
my free time with gEDA admin tasks and the rest (much smaller amount)
of my time hacking code. :( 


>Excuse me if I bring up old issues here - Google does not seem to 
>index the lists (at least the user list which I tried to search).


	Google should be indexing the lists now, but google's index
might be a little dated.  Either way, these issue are good issue to
be discussed.

								-Ales