Git Chronicles, updated?

Git Chronicles, updated?

[Jakub Narebski]

At "Git Together '08" Junio C Hamano has delivered "Git Chronicle, 
Recent Additions to Git" talk[1].  What's more Junio posted slides
for this talk[2] at http://userweb.kernel.org/~junio/200810-Chron.pdf
so that people who were not present at Git Together could read it too.

Now that's Git Together '10 comes near, I wonder if it would be possible
to have updated "Git Chronicles".  It was two years of development ago.
Hopefuly Junio still has tools he used to generate data for this talk.


One of graphs shown was growth of git codebase and of git contributors.
Did git development stabilized during those two years since 2008, or
does it still reads as active rather than stable development?

Another interesting graphs was plot showing number of surviving lines 
added in a give release relative to mumber of all lines added in said 
release.  This was used to detect which releases were important ones.
Were there any releases between 2008 and 2010 of significant importance?

The slides for "Git Chronicles" from 2008 closed with timelines of git 
features.  Were there any important user-visible features added since 
2008 (notes, sparse checkout, "smart" HTTP)?


What might be also interesting is a descriotpion of how some important 
feature came into being, with hint of an idea, discussions, prototypes, 
failed starts, dropped patches, reworkings, accepted version and then
improvements.  If one is not watching git mailing list regularly for a 
longer time, what one sees is the final product.  One doesn't know what 
it might to take to get a large feature into git...


[1] https://git.wiki.kernel.org/index.php/GitTogether08
[2] http://gitster.livejournal.com/17411.html 
[3] http://sourceforge.net/projects/ohcount/

-- 
Jakub Narębski
Poland

Re: Git Chronicles, updated?

[Junio C Hamano]

Jakub Narebski <jnareb@gmail.com> writes:

> Now that's Git Together '10 comes near, I wonder if it would be possible
> to have updated "Git Chronicles".  It was two years of development ago.
> Hopefuly Junio still has tools he used to generate data for this talk.

The tools are unfortunately very small parts of the story.

> One of graphs shown was growth of git codebase and of git contributors.
> Did git development stabilized during those two years since 2008, or
> does it still reads as active rather than stable development?
>
> Another interesting graphs was plot showing number of surviving lines 
> added in a give release relative to mumber of all lines added in said 
> release.  This was used to detect which releases were important ones.
> Were there any releases between 2008 and 2010 of significant importance?

These two are both interesting questions to ask, and luckily they are
mostly mechanical.  I'll try to dig up the tools to help generate them.

> The slides for "Git Chronicles" from 2008 closed with timelines of git 
> features.  Were there any important user-visible features added since 
> 2008 (notes, sparse checkout, "smart" HTTP)?

This was produced manually, out of release notes, blame output from the
Documentation/, as "feature dates" cannot be discerned mechanically
(e.g. history of "blame annotate shootout" needs to be written by somebody
who can tell that the current "git blame" came from "git pickaxe"
experiment, and all three variants were written from the general algorithm
description without code I wrote a long time ago).

> What might be also interesting is a descriotpion of how some important 
> feature came into being, with hint of an idea, discussions, prototypes, 
> failed starts, dropped patches, reworkings, accepted version and then
> improvements.  If one is not watching git mailing list regularly for a 
> longer time, what one sees is the final product.  One doesn't know what 
> it might to take to get a large feature into git...

Yes, it would be very interesting.  Will you volunteer to be one of the
git chroniclers?  It was a lot of work back then, and I don't think I can
do this properly as a two-day hack-job by myself.