List of Authors in aboutdata.cpp

[Ingo Klöcker]

On Montag, 23. August 2021 22:25:27 CEST David Bryant wrote:
> On Monday, August 23, 2021 3:54:21 AM CDT Ingo Klöcker wrote:
> > Isnip the bit about which repo Bryant drew his data from]
> > 
> > I'm wondering whether you also included the recent commit history of the
> > many different repositories KMail and its important libraries were split
> > into?
> 
> No. I just used the "kmail" repo from GitLab.

Don't put too much work into this. The statistics will never be perfect and 
your statistics are a very good start.

> > According to the history of the old kdepim and kdepimlibs (sadly,
> > currently
> > only available on GitHub: https://github.com/KDE/kdepim and
> > .../kdepimlibs)
> > repositories, I have made 1300+ commits ...
> 
> Well, I had drawn up a report, which is atttached. But that only includes
> some 26,000 commits, from the "kmail" repo on GitLab. I cloned the github
> libraries mentioned above this mornng. There are about 96,000 commits in
> "kdepim", and 22,000 in "kdepimlibs". So that's ~350% more than I had
> included originally. *But I notice the github logs only include history
> through August, 2016.*

Yes, those two repositories were split into a whole bunch of separate new 
repositories in 2016 which are now mostly in the PIM group in GitLab.

> I'll summarize all the data sometime, when I figure
> out where to find it all. It's a bit of work, because many people have
> spelled their names two different ways, etc. And it will have to come from
> many different repos, apparently.

Yes, but I repeat: Don't put too much work into this.

> Just FYI, Ingo, I find 1,268 lines in those two github log files that say
> "Author: Ingo K...". Your last name is spelled two different ways. I
> suppose it has something to do with UTF-8, which didn't exist when the
> KMail project started.

Ironically, the umlaut in my last name triggered my first contributions to 
KMail. :-)

> If I just use Kate to search for "ingo " I get 1,352
> matches in the first file (kdepim) and 53 in the second (kdepimlibs). So
> far as I can see, you are the only Ingo mentioned in any of these files.
> There is a guy named Rich Birch <ringo at albumsnaps.com>, so the blank is
> important. I had some trouble with the search function in Kate ... it hangs
> fire sometimes, and I have to kick it in the pants to get it to search the
> whole file. It's sort of large (34 MiB all together).
> > [snip some stuff ...]
> > Without having checked the actual contributions of those 11 authors, it is
> > not at all weird, that those 11 people are on the list of authors because,
> > obviously, those people are the authors of patches that have been
> > committed. There is a reason why git differentiates between authors (those
> > who actually write the code) and committers (those who merely add the code
> > to the source code repository).
> 
> That's not how I interpret the log file, Ingo. I see lines like this:
> commit b127ab527214b5e24a25dba737c0add4665d65f6
> Author: Ingo Klöcker <kloecker at kde.org>
> Date:   Sun Jan 9 00:44:56 2005 +0000
> 
> 
> So it appears that every commit has an Author, while patches show up like
> this: Patch by Sandro Knauss

Yes, every commit (in git) has an Author. All commits also have a Committer, 
but that's not shown by default. With git some patches commit by others have 
the correct author, i.e. the author of the patch. Older commits like the one 
above that predates KDE's move from Subversion to git have identical author 
and committer because Subversion didn't know anything about committers. 
Therefore we put a manual attribution "Patch by ..." into the commit log 
message.

> I understand your point, that writing the code to make a patch is usually
> harder than performing the purely clerical function of adding the new code
> to a repository. But most of the time people commit patches they have
> authored themselves. For example, there are 97,679 commits in the kdepim
> repo, while the word "patch " only occurs 1,599 times in the entire log
> file.

Sure. After a new contributor has sent in a few patches, they quickly get the 
permission to commit themselves. 

> And the majority of the occurrences of "patch " appear to be used in
> the commit message as a description of what the committer is doing, not to
> attribute authorship to somebody else. Compiling these statistics is hard
> enough if I rely on git to define what "Author" means. I might be able to
> read and interpret 100,000 commit messages before I die. But I don't want
> to go there.

Yes, what I wrote above. I suggest that you simply ignore the older history 
(before the split) and concentrate on the younger history. This may leave a 
gap, but it will include all contributions in the last few years which, in my 
opinion, is good enough for identifying contributors who should be added to 
the About dialog.

> > > [Monte] Indeed this dialog box is very big as there is a lot of people.
> > 
> > [Ingo] I don't see a problem with a long list of authors. Quite the
> > contrary.
> > 
> > > [Montel] I am agree that we need to reduce it (Perhaps we can add a
> > > specific file in repo to move name of people which made less than x
> > > commits.> 
> > -1
> > 
> > [Ingo] I very much disagree with removing any names from the list of
> > authors. Who are we to decide that people who have been added to the list
> > of authors 20 years ago are not worth anymore to be listed?
> 
> I don't necessarily want to remove any names. I want to (a) recognize people
> who joined the project fewer than 15 years ago and (b) reconcile
> discrepanciies between the credits in the "Handbook" and the list that pops
> up in Help --> About Kmail. I'm new here, and it looks to me as if nobody
> has paid much attention to this in a long time. I really just want to be
> fair to everybody who has had a hand in the project, to the extent that is
> feasible.

I understand, but I'd focus on the last 5 years. Apparently, almost all new 
contributors in the last 15 years or so didn't really care for being mentioned 
in the About dialog. Otherwise, they would have asked the maintainer to add 
them. Therefore, I think it's okay to accept a gap between 2005 and 2016 and 
leave filling this gap to another time resp. use the statistics that you have 
already gathered to cover those years.

> > For what it's worth, when I was maintainer of KMail in the early 2000s my
> > rule was to add anybody to the list of authors who made a non-trivial
> > contribution. One of my thoughts/hopes back then was that new contributors
> > who are given credits in the About dialog are motivated to stay longer and
> > contribute more. I think that a rule like "Your contributions are not
> > worth
> > adding your name to the About dialog until you have made 25+ commits."
> > will
> > motivate nobody to stay longer and make more contributions.
> 
> I'm not going to argue the point, Ingo. But nobody's name has been added to
> the "About" dialog since 2008, or 2009.

Well, apparently my successors didn't care to maintain the list of 
contributors in the About dialog. 

> Here, I've rambled on long enough. The following report was drawn from the
> "kmail" repo on Gitlab. Log file runs through August 20, 2021. People
> currently listed as "Authors" are listed *in red*, and "*Thanks to*" are in
> light blue. I'll update this list once I figure out how to obtain complete
> git log data.

I'm really grateful for your effort. I'll repeat myself once again: Don't put 
too much work into this. (But don't stop until you are satisfied with the 
result yourself.)

The following repositories are the most important ones that make up KMail:
https://invent.kde.org/pim/kmail
https://invent.kde.org/pim/messagelib
https://invent.kde.org/pim/mailcommon
https://invent.kde.org/pim/mbox-importer
https://invent.kde.org/pim/akonadi-mime
https://invent.kde.org/pim/kimap
https://invent.kde.org/pim/kmailtransport
https://invent.kde.org/pim/ksmtp
https://invent.kde.org/pim/kmbox
https://invent.kde.org/pim/kmime
https://invent.kde.org/pim/mailimporter
https://invent.kde.org/pim/kmail-account-wizard
https://invent.kde.org/pim/pim-sieve-editor
https://invent.kde.org/pim/libksieve

The following repositories also make up important parts of KMail, but they 
also contain parts of other PIM application like the address book and the 
calendar:
https://invent.kde.org/pim/kdepim-runtime
https://invent.kde.org/pim/kdepim-addons
https://invent.kde.org/pim/akonadi
https://invent.kde.org/pim/akonadi-search
https://invent.kde.org/pim/pimcommon
https://invent.kde.org/pim/libkdepim
https://invent.kde.org/pim/kpimtextedit
https://invent.kde.org/pim/grantlee-editor
https://invent.kde.org/pim/grantleetheme
https://invent.kde.org/pim/akonadi-airsync
https://invent.kde.org/pim/ktnef
https://invent.kde.org/pim/libkgapi
https://invent.kde.org/pim/libkleo

One cannot draw a clear line between those two groups. I put those 
repositories into the first group which I think are almost exclusively used by 
KMail. I'm certain that others would have grouped them slightly different and 
would have included repos that I didn't include and would have left out repos 
that I did include.

Regards,
Ingo
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/kde-pim/attachments/20210823/d05aaa36/attachment.sig>