2024-03-15 00:51:26 <KipIngram   > Happy pi day, btw.
2024-03-15 00:57:40 <thrig       > half tau
2024-03-15 05:09:45 <KipIngram   > So June 28 is Tau Day, then?
2024-03-15 14:27:05 <veltas      > zelgomer: Does this count as hoops? https://pastebin.com/jEVhLH2D
2024-03-15 14:27:34 <veltas      > For rare situation I'd want to do that, that will do.  I suppose this is what I mean by 'C is good enough', it's not perfect but it can get everything done.
2024-03-15 17:50:41 <veltas      > Got an idea for a word you can run after a colon def 'WEAK' that removes the definition if it's a redefinition
2024-03-15 17:51:09 <veltas      > Can write specialised/optimised definitions for a platform and then the WEAK cross-platform definition later
2024-03-15 17:53:18 <GeDaMo      > Couldn't you use [defined] or [undefined] for that?
2024-03-15 17:53:32 <zelgomer    > veltas: yes, that's the sort of hoop i mean
2024-03-15 18:07:51 <veltas      > GeDaMo: Yes but what's nicer?   [UNDEFINED] THING [IF] : THING .... ; [THEN]
2024-03-15 18:07:53 <veltas      > or  : THING .... ; WEAK
2024-03-15 18:10:26 <zelgomer    > veltas: one thing i'd consider is how it interacts with vocabularies. e.g., if it finds a previous definition, but it's in a different vocabulary than CURRENT, then should it copy an alias to CURRENT?
2024-03-15 18:31:43 <KipIngram   > I don't think most systems check for that - they just make a
2024-03-15 18:31:47 <KipIngram   > new definition.
2024-03-15 18:31:54 <KipIngram   > It's an interesting idea, though.
2024-03-15 18:33:31 <KipIngram   > You usually can have multiple vocabularies in your search path, though, so that list would guide you to the desired definition.  I'm not sure an alias would be useful, unless you wanted to "freeze in" the compile time selection of that original word.
2024-03-15 18:38:03 <zelgomer    > the purpose of the alias is that when you write  : thing ... ; weak  then there is an expectation going forward that "thing" will be reachable from that vocabulary
2024-03-15 18:38:44 <zelgomer    > the only way to keep that consistent is to, if "thing" is found in another vocabulary, create an alias header in the current vocab
2024-03-15 18:39:37 <zelgomer    > i would be pissed off with a vocabulary that may or may not have words defined in it depending on the context when the vocabulary was parsed
2024-03-15 18:41:12 <thrig       > usually those go by interfaces or abstrations you need to fill in elsewhere (or else)
2024-03-15 18:42:05 <KipIngram   > Ok.  It's not how I think of it, but there's nothing wrong with it.  I just figure it's up to me to arrange my vocabularies appropriately for what I'm doing.  But I pretty much never reuse names unless they're "local helper" names, in which case I want a new definition.  Or maybe I have a vocabulary for a certain "type" of data that offers a standard set of words - different types would call for different
2024-03-15 18:42:07 <KipIngram   > implementations.  I'd just choose the one I needed in some situation.
2024-03-15 18:43:43 <zelgomer    > KipIngram: you're missing the point. what veltas proposes means that the word /may not be defined in the current vocabulary at parse time/ depending on the context. in that case, how can you reliably arrange your vocabularies appropriately for what you're doing unless there is a promise that that word will exist in that vocabulary?
2024-03-15 19:21:42 <crc         > http://sprunge.us/ybDfXg would be a quick implementation of veltas' WEAK in one of my systems
2024-03-15 19:35:30 <zelgomer    > crc: what is sip?
2024-03-15 19:38:27 <crc         > &~scan sip   would be something like  DUP >R ['] ~scan EXECUTE R>  in a traditional forth
2024-03-15 19:39:52 <crc         > or, perhaps   : sip ( value address -- value )  OVER >R EXECUTE R> ;
2024-03-15 19:41:51 <zelgomer    > ok, like factor's "keep"
2024-03-15 19:56:16 <crc         > yes, looks like it. I think I took the sip name from something joy related