Sizes of TL distribution

TeX Live needs to have a small basic config + curated collections 
(e.g. Core Latex Collection) that can be easily added and installed.

A small core install that resolves without dependency hell.  A two-tier 
solution to leverage what Frank described.

Packaging standalone collections with locked deps, users get instant 
gratification + extensibility.

------------

Kristof Đymek ([email protected])
Engineer General - GNU Project

Society has become too quick to pass judgement and declare someone
Persona Non-Grata, the most extreme form of censure a country can
bestow.

In a new era of destructive authoritarianism, I support Richard
Stallman.  Times of great crisis are also times of great
opportunity.  I call upon you to make this struggle yours as well !

https://www.gnu.org     https://www.fsf.org/


> Sent: Wednesday, January 21, 2026 at 9:45 AM
> From: "Frank Mittelbach" <[email protected]>
> To: [email protected]
> Subject: Re: Sizes of TL distribution
>
> Am 20.01.26 um 11:32 schrieb Bruno Voisin via tex-live:
> > This probably will never happen, too much work involved, and not really 
> > in line with the open source philosophy, but still a dream. And, to be 
> > fair, the TLC already provides most of what's needed, directing the user 
> > to the relevant packages and describing their most useful commands.
> 
> well I think that dream is in line with open source philosophy, after 
> all nothing prevents new packages from appearing or existing being used, 
> but I agree with you that once in a while trimming the forest helps to 
> actually grow it again. And this is precisely what is in parts happening 
> with the tagging project as we need to do that kind of streamlining to 
> make it work in the end. So packages vanish (or rather move in one way 
> or the other into the kernel and the original will be emulated for 
> backward compatibility). Last not least quite a number of package that 
> have gone stale with no maintainer and bugs accumulating and it is 
> difficult for ordinary users to understand that what was recommended in 
> a post somewhere 10 year back may no longer be correct because the 
> package didn't adapt (quite true for a lot of LLM advice too).
> 
> But I fear nothing in that direction really makes the distribution size 
> much smaller, after all LaTeX is quite committed to to backward 
> compatibility so even if, say, enumitem would no longer be necessary to 
> load, one would still need something with that name for old documents or 
> for those interfaces we decided not to provide, etc.
> 
> And then there are all the special packages, that are either totally 
> niche or just yet another variant of the same scheme but dear to some 
> people or used in old documents ...
> 
> However, yes with the TLCs we tried to provide a good foundation that 
> gets the job done for many people and many tasks. And while I may have 
> made the wrong judgement call in one or the other place I think on the 
> whole it came out well, certainly TLC1 and 2 and TLC3 has to still prove 
> itself a bit (but there I'm also fairly happy about the result).
> 
> So TLC3 as a scheme wouldn't be too bad an offer in my opinion (biased 
> as I am).
> 
> frank
>