rename rules -> op_rules
Really, *all* of our files contain proof rules
-
Maintainer
But op rules makes even less sense... Then call it wp_rules?
-
Maintainer
Now it looks like this file is specific to bin_op and un_op or so.
-
Author Owner
But op rules makes even less sense...
It certainly makes more sense than just "rules". I agree it may not be perfect.
Then call it wp_rules?
Containing rules for WP is hardly a distinguishing feature of this file. What about lang_wp or wp_lang?
-
Maintainer
I think rules covers it fairly well, this file contains the core proof rules of the language. All the other stuff is just derived on top of that.
But since you don't like it, how about just
weakestpre
by analogy toprogram_logic/weakestpre
. Or alternatively,base_rules
. -
Maintainer
It is in the root of the folder
heap_lang
, that should make that clear. -
Maintainer
First of all, @jung, I don't quite see how
op_rules
makes more sense thanrules
. I don't see whatop
refers to : operational? operators? Then, I don't likewp_lang
orlang_wp
, becauseheap_lang/lang_wp
is redundant.So, now, I like
weakestpre
orwp
. We could even go back tolifting
, which, I think, was quite a good name. -
Maintainer
JH precisely describes my concerns. We then also have to prefix all the other files with
lang_
.I am not so sure whether
lifting
still makes sense,program_logic/lifting
contains the generic lifting lemmas/rules, whereas this file just makes use of it. -
Maintainer
I am not so sure whether lifting still makes sense, program_logic/lifting contains the generic lifting lemmas/rules, whereas this file just makes use of it.
That is the point : in a sense, this file specializes the lifting lemmas to this language.
-
Author Owner
op
is for operational. But okay, if you both don't like it, then let's look for sth. else.heap_lang
contains lots of things related to the language, the primitive language-specific triples are fairly particular. I think that justifies havinglang
twice, like it is twice inheap_lang.lang
. -
Maintainer
op is for operational. But okay, if you both don't like it, then let's look for sth. else.
Ok. But the point is that here we are not describing the operational semantics, but rather an axiomatic semantic.
rules_op
would probably be more appropriate for the currentlang.v
.heap_lang contains lots of things related to the language, the primitive language-specific triples are fairly particular. I think that justifies having lang twice, like it is twice in heap_lang.lang.
I am not too strongly opposed to
wp_lang
. Enven thoughweakestpre
orlifting
would make more sens in my opinion. -
Maintainer
My preferences:
- rules
- weakestpre
- lifting
- lang_rules
- lang_weakestpre
I will veto against rules_op.
-
Author Owner
- lang_wp
- lang_weakestpre
- lifting
- {base,basic}_{wp,weakestpre}
I will veto anything involving "rules" (I am aware this includes vetoing the current name; the reason I used it is that I wanted to stick to the term you used but make it more descriptive). I could maybe accept just "weakestpre", but I still think it is a horribly indescriptive name.
-
Maintainer
Alright, then let's go back to
lifting
as the file was called before. JH seems to prefer that name, and it is the first common name on both of our lists. -
Author Owner
Ack. Will do.