Re: gEDA-dev: SoC Hopeful

[DJ Delorie <dj-at-NOSPAMdelorie_dot_com>]

> Is it possible to decide beforehand upon a fixed order of columns
> for the csv files involved.  This would allow for easy merging of
> contribs by other users/developers/maintainers.

I've never considered that.  I usually reserve the first row of csv
files to be the headers, and merge the columns by name.  Also, the csv
parser on gedasymbols allows for pre-filled columns (i.e.  columns
where every cell has the same value) via named attributes in a prolog.

> No mentioning of gnetlist backend, pcb HID or gnucap specific fields so
> far.

No, but I suspect there will be.

> I could imagine that a certain part/package can be sold by different
> vendors with different prices (due to currency and/or location as in
> Europe, USA, Asia etc.)

Hence the separate vendor databases.  I buy from both Mouser and
Digikey, with each part chosen by price.  Currently, I have a perl
script that scans some text files to figure out which part comes from
which vendor.

> IMHO a third database file/table is required for local determined
> entries (vendor, currency, price etc.).  I could think of
> "vendors_and_prices_NL-nl.csv" localisation scheme.

Perhaps.  I suspect the legal issues behind storing other vendors'
prices in a public file will preclude us worrying about this.  Each
site would have their local vendor price files, the public files just
match manuf. numbers to vendor numbers.

> What is the use of the "author" entry in the first database file/table.
>
> What if you happen to have an author-1 for the footprint and an author-2
> for the symbol and an author-3 for the gnucap model etc.

In the gedasymbols case, I don't expect people to be mixing symbols and
footprints that way.  A csv would be ONE of the following:

* symbol, description, copyright, author
* footprint, description, copyright, author
* symbol, footprint, part numbers

Alternately, a symbol reference file could suggest footprints yet
still have the symbols' authors.  I don't expect to try to get a
symbol's author and a footprint's author in the same row.

I.e. I'm trying to reuse the file format for multiple purposes, and
relying on cross referencing (cross table queries in SQL) to build the
rows I need from the pertinent information.  I.e. if I'm building a
BOM I don't care who the symbol authors are, so I don't read that
table/column when I merge all the data together.


_______________________________________________
geda-dev mailing list
geda-dev@moria.seul.org
http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev