RE: [vhdl-200x] Issue with VHDL name attributes

Jim,

This may be moot, but for the record:

My comment was not intended to suggest in any way at all that the ability to have multiple architectures should be removed. (I've been a strong proponent of this feature since we defined in in 1984 or so.) My suggestion was only intended to allow (not require) the definition of an 'anonymous' architecture, which effectively has either no name, or the name of the entity, depending upon how you look at it.

I suggested this because the only reason I can see for wanting to name the architecture the same as the entity is if there is only one architecture, and therefore there is no need to have different names for different architectures. In particular, examples in the LRM typically have only one architecture.

Regards,

Erich

| -----Original Message-----
| From: Jim Goeke [mailto:james.goeke@kodak.com]
| Sent: Thursday, April 08, 2004 6:27 AM
| To: peter@ashenden.com.au; johnr@model.com;
| SBailey@model.com; Erich Marschner
| Cc: vhpi-pilot@eda.org; vhdl-200x@eda.org
| Subject: RE: [vhdl-200x] Issue with VHDL name attributes
|
|
|
| Please see my notes below.
|
| Regards,
| James Goeke
|
|
| >
| >Peter,
| >
| >
| ...
| >
| >Architectures (or "architecture bodies", as they were originally
| >refered to)
| are to entity (interface) declarations as package bodies are
| to package
| declarations. The only significant difference w.r.t. scope
| and visibility is
| that we wanted to allow alternative architectures, whereas
| one package body was
| deemed sufficient. The need to allow multiple architectures led to a
| requirement for names on the architectures.
| >
|
| ++++
| The use of multiple/different architectures for a single
| entity is a very
| powerful and useful tool. An example of where we use this
| frequently is
| embedded memory blocks. We will usually have a very high
| level behavorial model
| that we write that is optimized for speed. We will have
| another architecture
| that is the vendor simulation model. Often there is an order
| of magnitude
| difference in simulation speed. Compiling two architectures
| allows extremely
| simple switching between the two, or even simultaneous access
| for different
| simulations.
| ++++
|
| >
| >If the goal was to allow an architecture to have the same name as the
| corresponding entity, a much more natural solution would have
| been to simply
| make the architecture name optional:
| >
| >entity E is ...
| >end;
| >
| >architecture of E is ...
| >end;
| >
| >There is only one name, and it is shared by the entity and the
| >architecture -
| which together still constitute a single declarative region.
| >
|
| ++++
| This would appear to me to break the ability to have multiple
| architectures. If
| that is true please do not do it. Removing the capability to
| use multiple
| architectures for name scoping rules is a bad trade off in my opinion.
| ++++
|
| >
| >Regards,
| >
| >Erich
| >
| >
|
|
Received on Thu Apr 8 06:23:17 2004