On Mon, 11 Feb 2008, Bill Page wrote:
| On 2/11/08, Gabriel Dos Reis wrote:
| > On Mon, 11 Feb 2008, Bill Page wrote:
| > |
| > | As I said before, allowing arbitrary strings of symbols to be a member
| > | of Symbol could be very similar to how subscripts are represented.
| > |
| > | (5) -> script(alpha,[[i,j],[k,l,m],[a],[b,c]])::OUTFORM::SEX
| > |
| > | (5) (*02312alpha i j k l m a b c)
| > | Type: SExpression
| >
| >
| > Am I the only one to be bothered by this purely syntactic encoding
| > of something that could have semantics content?
| >
|
| If you are referring to the name-mangling to encode subscript
| information, then no, you are not the only one. Actually, I think it
| is rather shocking to see such a hack here! Surely there is a better
| way. It would be interesting to see what breaks if one attempted to
| replace the internal representation of the Symbol domain with
| something more "axiomatic".
I came to this about a year ago, when I was trying to write some
algebras based on `Symmetries and Differential Equations' by Bluman&Kumei.
I wanted to abstract over the scripts, only to discover that the
parser has its own idea about scripts.
| > I've always been annoyed that the parser would decide to encore
| > subscript, before semantics processing starts.
| >
|
| It is not clear to me what part the parser plays in this, nor how this
| might differ between the Axiom interpreter and Spad. Perhaps you can
| explain?
See src/interp/postpar.boot
postScriptsForm(["Scripts",op,a],argl) ==
[getScriptName(op,a,#argl),:postTranScripts a,:argl]
postScripts ["Scripts",op,a] ==
[getScriptName(op,a,0),:postTranScripts a]
getScriptName(op,a,numberOfFunctionalArgs) ==
if null IDENTP op then
postError ['" ",op,'" cannot have scripts"]
INTERNL("*",STRINGIMAGE numberOfFunctionalArgs,
decodeScripts a,PNAME op)
`postparsing' is done long before the compiler does anything. And
this is not the only place where unfortunate semantics decisions are
made by parser long before we get to semantics processing.
| But it seems obvious that the interface to the Symbol domain is
| intended to serve rather general purposes. Consider for example the
| definition:
|
| elt(sy,lx) == subscript(sy,lx)
|
| which seems to imply that writing
|
| A[i,j]
|
| should result in a call to subscript$Symbol. Writing:
|
| (1) -> )set mes bot on
| (1) -> A[i,j]
|
| in the interpreter confirms this. So I think one might claim that the
| decision to treat 'A[i,j]' as a subscripted symbol is in fact a
| semantic one. No?
No, you're seeing only half of the picture :-)
See above.
-- Gaby
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
open-axiom-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/open-axiom-devel
