Re: e2sr name mangling

[mcn@aps.anl.gov (Matthew Needes (LANL))]

>RE: Matt Needes proposed changes to e2sr:
>
>I believe the whole approach of extensive use of properties and substitution
>in the schematic itself is clumsy, in that it requires additional work to
>maintain the schematic.  At CEBAF, we have used an additional utility program
>
>which encapsulates our desired naming convensions into rules, which operate
>correctly (99.9% of the time) on any record anywhere in the hierarchy, so that
>there is no burden to editing the schematic to get the names right.
>
>I think it would be a good idea to merge the two capabilities of e2sr and
>our name mangling program, but it should not be dependent upon setting
>properties in each record, but should instead be completely rule based.
>
>Chip

Part of the advantage of having more information on the schematic is that
you have more information at your fingertips while browsing a design.

A flat list of hierarchical substitution rules may or may not make much
sense until you compare them with your design anyway.  Why not just put
the rules on the schematic ?

Perhaps the additional schematic maintenance is not a big deal, because
in the absence of other requirements, a consistent naming scheme should
be the rule rather than the exception.  If you keep to a particular
standard you will avoid making wholesale changes to record instance
properties.  You may, however, have to maintain set properties on hierarchical
symbols (as in the PV substitutions).  Again, I think that watching these
values as you run down your design is an advantage.

If you want to define a different standard for yourself, you may have
to modify your symbol library (or perhaps just edb.def), and perhaps provide
"rules" (variable substitutions) at the top level of your schematic.

While I have not investigated the full power of either approach, I think
that top-level hierarchical substitution rules are equivalent to textfile
based substitutions anyway.

Matt