[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: gEDA-user: Introduction and some questions/thoughts on gEDA/gaf...



> 
> [snip]
> >Sorting chips by function (opamp, regulators, microcontrollers..) should be 
> >better since manufacturers of some chips change but the names are preserved.
> >For example old uA723 is now made by Texas and not by Fairchild who introduced
> >the prefix uA (originally "micro"A) in 1970's. Some popular chips can have 
> >many manufacturers, so it will make the problem where to put them.
> >
> 
> 	Hmmm.  An interesting idea.  Currently, there are directories 
> (like micro or tube) which are exactly this.  Does this really scale?
> What do others think of this?

Designing an optimal parts library is more difficult than one
might think at first blush.   On one hand, there are lots of parts
which aren't associated with any one particular manufacturer, such as
passives, common diodes (1N4148), transistors (2N2222)  or (showing my
age) many common vacuum tubes.    On the other hand, it's logical to
organize other parts by manufacturer.  Examples which spring to mind 
are MicroChip PICs, Intel microprocessors, or speciality parts such as
the MAX232.  

The corporate parts databases I have seen usually group parts by
functionality or class, not by manufacturer.  That is, the grouping
might be (for example), resistors, caps, logic ics, discrete
semiconductors, analog ics, power ics, etc.  The problem with this
scheme -- besides the fact that it is usually poorly implemented -- is
that lots of parts fall between the cracks.  Where do you put A/Ds or
D/As in this scheme?  In analog ics or digital ics?  Yes, you can
create a new "conversion ic" catagory, but soon you have more
catagories than Carter has pills.  And it always seems like there is a
missing catagory no matter how many catagories you already have..   

My opinion is this:  Ultimately, all ways of organizing the parts
database are equally bad.  Therefore, a better solution would be to
create a better parts browser.  Ideally, a parts browser would
incorporate a Google-like search mechanism allowing you to type in an
aribtrary string -- e.g. component part no, some component properties
("resistor 1.00K 0805"), or company name and functionality ("analog
devices opamp") -- and let the search engine give you a list of
possible parts from which to choose.  In this scheme, the parts could
be organized in any way desired (by mfr, by catagory, or whatever),
and you could either find them the usual way, or find them by using a
Google-type string search.

I realize that this is an ambitious change to gEDA's current parts
browser, so it may not happen any time soon.  Alas!  We can dream,
can't we?!?  Meanwhile, I think we could simply tweak the current,
mixed scheme, and it would be a worthwhile improvement.  

Stuart