On Monday, 31 March 2025 at 17:38:06 UTC-7 Trevor Karn wrote:

What was the original intent behind having the dual requirements of (i) a 
system python and (ii) a SPKG python both in Sage? What (once upon a time) 
did having a SKPG do that couldn't/shouldn't/wouldn't be done by the system 
python?


At the very start of sage (some 20 years ago) a big selling point of sage 
was to bundle all software necessary for sage *together* with tight version 
specs set between the components (really: fixed version combination, 
because everything was bundled). At the time this was great, because open 
source math software was notoriously difficult to get built. Sage was 
actually a fairly reliable way of getting a bunch of mathematical software 
actually built and usable on your machine. (Sage was about "building the 
car, not reinvent the wheel", so most functionality was in libraries and 
systems shipped, with the sage library primarily providing glue to easily 
commicate with the different component. So in that sense, sage really 
started out as a distribution)

Python wasn't as standard yet either, so a choice was made to have *some* 
python requirement to get things bootstrapped (I guess to avoid having to 
write shell scripts that would have to deal with all the subtle 
incompatibilities between different shells), but then build python in order 
to ensure we knew exactly what version to rely on for actual operation. At 
some point, the sage python package did have some patches included.

Python has now become much more standard and probably a bit more stable in 
its features too. So it may well be that the benefit from having an exact 
python isn't as big any more. We're not usually building gcc either.

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion visit 
https://groups.google.com/d/msgid/sage-devel/c219eac9-1da4-43ba-a034-e86e62b12d57n%40googlegroups.com.

Reply via email to