Mail Archives: geda-user/2015/09/11/17:04:38
On Sep 11, 2015, at 2:38 PM, DJ Delorie <dj AT delorie DOT com> wrote:
>
>> Even though you have to remember attributes and values yourself it
>> is not just arbitrary text. For already defined attributes it would
>> have been nice with a list to choose from and for each attribute a
>> list of predefined values to chose from where appropriate.
>
> It's not "just arbitrary" because some other tool (typically gnetlist
> for gschem) interprets the values. But as far as gschem itself is
> concerned, aside from a few convenience pulldowns that provide
> commonly used defaults, the text really is just arbitrary.
>
> If you were using gschem with something other than gnetlist, which had
> a different interpretation of the text, it would still not be "just
> arbitrary" to that other something, but not because gschem itself gave
> it any meaning.
For the majority of attributes, (value, footprint, pintype, …) it isn’t really gnetlist but the back end that interprets. Nice factoring, really.
>
> Thus, my position that adding arbitrary text *in gschem* is not
> difficult.
Sure.
>
> The task is more difficult in something like pcb because the
> attributes have meaning *to pcb*, and pcb usually has to interpret
> them in realtime as the user is editing. It's the meaning that causes
> the difficulty, not the attribute storage itself. The realitime
> requirement often dictates a non-"arbitrary" internal storage as well.
That has more to do with the design of geda-pcb than with any real requirement. I don’t see why a geometry editor couldn’t edit a “teflon” or “constantan” layer. It isn’t the editing requirement that drives these restrictions. It’s the fact that geda-pcb doesn’t really model a PCB, but merely a restricted subset of PCB structures. And that also makes geda-pcb hard to understand for at least some of us.
>
> (this is similar to the XML "problem" - XML defines a structure for
> storing information, but doesn't give any meaning to the information,
> leading to lots of incompatible implementations)
>
It’s just mathematics, really. The meaning of the abstraction is separate from the properties of the abstraction. The application gives the abstraction meaning. Thus, Fourier analysis is useful for a very wide range of problems despite its origin in analyzing heat flow in rectangular objects. This is not a problem, but a powerful property of abstractions.
John Doty Noqsi Aerospace, Ltd.
http://www.noqsi.com/
jpd AT noqsi DOT com
- Raw text -