If "what's an author?" is the only issue we need to settle, that's
pretty easy.
An "author," it seems to me, is the person (or person) who did the
preponderance of the work initially, or as part of an enhancement. By
"the work," I mean the conception, design, coding, test-case writing
and debugging. That rules out bug-fixing, unless the bug was so gnarly
that it necessitated a gut-rewrite or major refactoring. It also rules
out Javadoc tweaks.
For example: I did a bunch of re-writing on MailUtil about 6 months
ago that added JNDI features, completely re-factored the main
processing methods, and added several private functions. I changed,
touched, and re-wrote 80% of the code. For that reason, I added myself
as an author.
But just as you have, I've also touched many many classes to fix bug
here and there. I can't think of a single case where it seemed
appropriate to add myself as an author.
Thus: if you design, write/enhance code, and write unit tests, you
might be an author. If you fix a bug, you probably aren't.
Andrew
On Mar 31, 2008, at 2:58 PM, Janne Jalkanen wrote:
But again, how to treat patch contributions? Whoever applies the
patch decides whether to take in the @author-tags? Where do we draw
a line if someone sends on a three-line patch, one of which is the
@author?
Apache is a high-profile society, and people will want to have their
name associated with it.
/Janne
On 31 Mar 2008, at 21:33, Andrew Jaquith wrote:
Folks,
The changes we're talking about beg the question: "what's the
problem we are trying to solve?"
In an earlier e-mail, I noted that the "we need to do what Apache
does" argument is bogus because Apache projects DO use @author
tags... or putting it another way, they don't remove them
consistently. I'm happy to furnish dozens, if not hundreds, of
examples of @author tags in the Tomcat project if you need further
convincing.
Equally bogus is the "the authors might get sued if we leave them
in" argument, because it ignores the reality of how modern e-
discovery is done. There are lots of ways to find people to sue,
for example trolling through the jspwiki-dev mailing list.
The "we are a community so we have to release community projects
without attribution" reason is NOT bogus, but it IS mighty
presumptive. I joined the JSPWiki dev team to have fun and make a
personal difference, not because I wanted to be part of Apache PER
SE (though I like that too).
Here's the bottom line. There are two geniune, defensible reasons
for keeping @author tags in the code:
- Personal attribution/pride
- Helping to establish expertise ("archaeology") for fixes, in
combination with commit tags
There are also two geniune, defensible reasons to remove them:
- When lineage of authorship is diffuse, or uncertain
- When an author does not want to attract unwanted e-mails about
the code
I seems to me that these reasons all point to personal decisions,
and should be made at time of commit. So I agree with Harry. I
don't think this is something we should legislate. The list of
authors is quite small, and the committers even smaller. It seems
to me a laissez-faire policy would work best.
Andrew
On Mar 31, 2008, at 2:00 PM, Dave Wolf wrote:
IMHO, if the rest of Apache projects don't use the author tags
(whether it
is enforced or not), then for consistency we shouldn't either. It
would be
confusing to new folks.
Additionally, (again opinion) an open source project should not be
tied to
individuals, so I agree with Janne that questions should go to the
lists. If
someone who has knowledge is still active then they can respond.
If the
author / knowledgeable person is no longer active, then the
absence of a
response might be informative to the wider community.
Dave
On Mon, Mar 31, 2008 at 11:50 AM, Harry Metske <[EMAIL PROTECTED]
>
wrote:
I think we shouldn't dictate right now, how to deal with those
conflicting
situations.
Is it very likely to occur ?
If it's not strictly necessary to make the choice now, I think we
should
let
it free.
But if anyone else on the list has a good reason to choose for an
all-on
or
all-off, please speak up.
regards,
Harry
2008/3/31, Janne Jalkanen <[EMAIL PROTECTED]>:
Then what do we do with new people if some of them insist to
have the
@author tag and some don't?
/Janne
On 31 Mar 2008, at 19:58, Harry Metske wrote:
Although I'm not a committer I'd like to comment :
I think both of you are right, there is code that is submitted
initially
fairly complete, and is not patched much afterwards, and if it is
patched,
it is only small bug fixes. In those cases you would want to
keep the
@author tags.
On the other hand there is code that is initially submitted as a
simple
first prototype, and is afterwards greatly enhance by others. In
those cases
you should remove the @author tags.
If it is not exactly mandated by ASF how to deal with, let
everyone
decide
for himself how to handle it, and you don't have to handle every
piece of
code the same, for some pieces keep your @author tag, for others
remove them
?
How about that, everybody happy ?
regards,
Harry
2008/3/31, Janne Jalkanen <[EMAIL PROTECTED]>:
Case in point: The author of LuceneSearchProvider has not
contributed anything in what, two years? Yet he still
stands in
as the @author. Would you go to him to ask for help, or
should
you go to the mailing list?
I'd ask the list, but I would likewise *know* that the author
hasn't
been seen for two years, information that would be lost if the
tag had
been removed.
The @author tag does not encode that information. Therefore,
nothing
would be lost, if it were removed (and it might probably be a
good
idea to remove it at that stage).
The code should really stand on its own legs.
You're saying you have your name on code you don't understand?
Yes. Other people have modified code for which I am still the
@author, and I can't really claim that I know what that code
does
anymore... And I'm pretty sure I have modified a lot of the
code
that other people claim to be @author of, and they no longer
have any
idea what is going on.
That's how it goes with projects that have been running for
nearly
seven years now.
/Janne
--
met vriendelijke groet,
Harry Metske
Telnr. +31-548-512395
Mobile +31-6-51898081
--
met vriendelijke groet,
Harry Metske
Telnr. +31-548-512395
Mobile +31-6-51898081
--
Dave Wolf
H: 303-377-9537
M: 303-956-9106
http://www.youtube.com/watch?v=EsMumUEPMCY Donate to VoteVets
"Our lives begin to end the day we become silent about things that
matter."
--Dr. Martin Luther King Jr.
