Part 2: Future of RAT

[Jochen Wiedmann <jo...@gmail.com>]

Hi,

now that the waves are flattened again, the first thing I should
probably do is to apologize: Had I anticipated the multitude of
reactions, I'd have put my question to this mailing list, rather than
to several. However, I expected no more than 2 or 3 replies, that's
why the broad audience seemed at the time to make sense.

Now, back to the rat-dev list, I'd like to put the same question to
you again: Isn't it time for RAT to leave the Incubator? And if so,
what would be the target?

Regardless of what was written in the first thread, I still feel that
the project is small in all aspects that I can think of. That's why I
clearly thought of becoming a subproject of another, like Commons. But
I did neither expect the pressure in the direction of becoming a TLP
(Greg Stein being interested), nor the lack of interest from the side
of Commons. (The other TLP's mentioned should have even less.
Personally I can't seem RAT related to Maven, or Ant. OTOH I do see a
relationship of RAT as a tool of general interest with Commons.)
Likewise, to me it seems misplaced in Infra.

But the alternative, remaining in the Incubator is something which I
consider to be worse in the medium term. To me, RAT is a project which
could have skipped the Incubator right from the start. The less it
should remain where it is.

But all of these are, of course, my personal impressions and perhaps
not shared by anyone else.

Should the discussion be heading into the direction of RAT as a TLP
(which I still view as the second best only) and the question of a
chair become a problem, then I'd offer myself - as the third or fourth
best option. Personally, I'd clearly prefer more experienced people
like Robert, Stefan, or Henri.

Jochen

-- 
I Am What I Am And That's All What I Yam (Popeye)

Re: Part 2: Future of RAT

[Stefan Bodewig <bo...@apache.org>]

On 2010-08-17, Ross Gardler wrote:

> The design of RAT makes it quite easy to do the very first of these
> (check for other headers and compare against a compatibility table),
> but even though I want it I've never found the time to do it.

It can already do that - well the Antlib can, so the core of RAT is
capable of doing it.  Just take a look at the javasources-w3c target in
ant-task-examples.xml, this will generate a report on RAT's source tree
where the "compatibility table" only contains the W3C license - flagging
all Apache licensed files as unknown.

For your usecase you'd add matchers for your compatibility table and all
files with a different license would fall out.

Stefan

Re: Part 2: Future of RAT

[Ross Gardler <rg...@apache.org>]

On 17/08/2010 12:56, Jochen Wiedmann wrote:
> On Tue, Aug 17, 2010 at 1:46 PM, Ross Gardler<rg...@apache.org>  wrote:
>
>> So for now I'd rather keep the question simple:
>>
>> "Is it done?"
>
> And that's exactly the question I cannot answer. You seem to have a
> bigger vision for what RAT *could* do. And I cannot share that vision,
> because I have no idea what that might mean technically.

OK, fair enough. I'll answer your question then.

For me an audit tool will assist an RM, at present RAT assists 
developers with license headers. So audit focussed features might include:

- identify (potentially) incompatible licenses in releases
- ensure notice files are complete
- ensure all copies of bundled libraries have their licences present
- ensure version numbers have been correctly updated
- check distributions are signed properly
- ???

Added "nice to have" extras that go beyond release audit could include:

- ensuring all developers are appropriately credited
- build change logs
- uploading verified distribution files
- test downloads have been mirrored correctly
- sending notification mails
- maybe even provide a full release management workflow tool

Obviously the list could go on forever and hence my question, is it done?

The design of RAT makes it quite easy to do the very first of these 
(check for other headers and compare against a compatibility table), but 
even though I want it I've never found the time to do it. Yet I do use RAT.

Hence my question "is it done?"

Personally I'd love to see RAT grow into a genuine Audit tool. Just this 
week I needed to use one because someone claimed one of my releases had 
a GPL file in it. It took 10 minutes with a real license audit tool [1] 
to tell me they were false positives from Ohloh.

Why not use that Audit tool?

All it does it check for licence headers. It's essentially RAT with a 
front end and more complete matchers. I would like to see more (see 
above) Flossology is GPL, hence I like the idea of RAT being a full 
license audit tool and possibly more.

However, as I said earlier, my wishes and desires don't write code ;-)

Ross

[1] http://www.fossology.org

Re: Part 2: Future of RAT

[Jochen Wiedmann <jo...@gmail.com>]

On Tue, Aug 17, 2010 at 1:46 PM, Ross Gardler <rg...@apache.org> wrote:

> So for now I'd rather keep the question simple:
>
> "Is it done?"

And that's exactly the question I cannot answer. You seem to have a
bigger vision for what RAT *could* do. And I cannot share that vision,
because I have no idea what that might mean technically.

I've been RM for a lot of Apache projects. I'd welcome ideas how to
improve the process. But I don't know what you expected from the
"original proposal". Hence no possibility for me to answer whether "it
is done".

Jochen

-- 
I Am What I Am And That's All What I Yam (Popeye)

Re: Part 2: Future of RAT

[Stefan Bodewig <bo...@apache.org>]

On 2010-08-17, Ross Gardler wrote:

> So for now I'd rather keep the question simple:

> "Is it done?"

For all I wanted RAT to do, it is.

That doesn't mean I wouldn't be around to help or even contribute as
time permits if people want to implement the full scope of the proposal,
I just lack the time and desire to drive it.

Stefan

Re: Part 2: Future of RAT

[Ross Gardler <rg...@apache.org>]

On 17/08/2010 11:10, Jochen Wiedmann wrote:
> On Tue, Aug 17, 2010 at 12:03 PM, Ross Gardler<rg...@apache.org>  wrote:
>
>> It's called Release Audit Tool and the original proposal was quite broad in
>> scope.
>
> I'm a plain stupid developer who has used RAT since quite some time
> without additional ambitions. Tell me what you'd expect, not what's
> written in a proposal, please.

This thread has taken a turn away from my original post. I did not say 
that I expect more of the RAT community, I asked is it done? I asked 
because the answer affects the future of RAT which is the topic of this 
thread.

It is clearly not done with respect to the original proposal, but it is 
clearly a useful tool. One that adds to the utility of the perl scripts 
that did exactly the same header checks before RAT exists.

I have no problem with people stating it is done and leaving it as is. 
What I want to know is "is that it?"

If it is then it is not an audit tool, it is a license header checker 
and we need to be careful about people thinking it is more.

Finally, I'm happy to answer the question of what I would like it to do. 
I hinted at a couple of issues that go beyond the original proposal in 
my first post but I don't want to derail a discussion about the future 
home of RAT with detailed technical discussion about what new features 
the tool might have.

So for now I'd rather keep the question simple:

"Is it done?"

Ross

Re: Part 2: Future of RAT

[Jochen Wiedmann <jo...@gmail.com>]

On Tue, Aug 17, 2010 at 12:03 PM, Ross Gardler <rg...@apache.org> wrote:

> It's called Release Audit Tool and the original proposal was quite broad in
> scope.

I'm a plain stupid developer who has used RAT since quite some time
without additional ambitions. Tell me what you'd expect, not what's
written in a proposal, please.

Jochen


-- 
I Am What I Am And That's All What I Yam (Popeye)

Re: Part 2: Future of RAT

[Ross Gardler <rg...@apache.org>]

On 17/08/2010 10:33, Jochen Wiedmann wrote:
> On Mon, Aug 16, 2010 at 11:36 PM, Ross Gardler<rg...@apache.org>  wrote:
>
>> So where is RAT going with respect to functionality. Is it really "complete" in it's current form, as some people have suggested?
>>
>> Don't get me wrong, it's a great tool for a very limited use case (I do use it outside the ASF), but a top level project that stops at that functionality is not likely to draw more patches from me since there are other more complete tools that fill my wider audit needs (albeit not Apache licensed).
>
> AFAIK, you're the first one to mention such thoughts. At least, I am
> unaware of any feature requests or something similar.

It's called Release Audit Tool and the original proposal was quite broad 
in scope.

Ross

-- 
rgardler@apache.org
@rgardler

Re: Part 2: Future of RAT

[Jochen Wiedmann <jo...@gmail.com>]

On Mon, Aug 16, 2010 at 11:36 PM, Ross Gardler <rg...@apache.org> wrote:

> So where is RAT going with respect to functionality. Is it really "complete" in it's current form, as some people have suggested?
>
> Don't get me wrong, it's a great tool for a very limited use case (I do use it outside the ASF), but a top level project that stops at that functionality is not likely to draw more patches from me since there are other more complete tools that fill my wider audit needs (albeit not Apache licensed).

AFAIK, you're the first one to mention such thoughts. At least, I am
unaware of any feature requests or something similar.

I'd be interested in that discussion, but it is unclear to me, whether
we are opening a second discussion here or whether this actually is
part of the question, whether, when and how to leave the Incubator.

Jochen

-- 
I Am What I Am And That's All What I Yam (Popeye)

Re: Committers (was Re: Part 2: Future of RAT)

[Ross Gardler <rg...@apache.org>]

On 17 Aug 2010, at 06:10, Stefan Bodewig <bo...@apache.org> wrote:

> [splitting this part out of the thread because it would distract from
> the rest of my response in a different email]
> 
> On 2010-08-16, Ross Gardler wrote:
> 
>> Although note I'm only a mentor and early contributor, not a
>> committed.
> 
> You are one of the four committers who have been voted in last December,
> together with Gav, Niall and David.
> 
> Stefan

Really? Oh that makes me a pretty crap mentor then ;-)

Sent from my mobile device.

Committers (was Re: Part 2: Future of RAT)

[Stefan Bodewig <bo...@apache.org>]

[splitting this part out of the thread because it would distract from
the rest of my response in a different email]

On 2010-08-16, Ross Gardler wrote:

> Although note I'm only a mentor and early contributor, not a
> committed.

You are one of the four committers who have been voted in last December,
together with Gav, Niall and David.

Stefan

Re: Part 2: Future of RAT

[Ross Gardler <rg...@apache.org>]

On 17/08/2010 14:42, Hyrum K. Wright wrote:
> On Tue, Aug 17, 2010 at 2:55 AM, Ross Gardler<rg...@apache.org>  wrote:
>> Something where I could say "are we overlooking a potential committer" or "do any committers appear to have gone emeritus without telling us".
>
> You may be interested in something we've used with Subversion for
> several years: the contribulyzer.  It looks at log messages (which
> must have a special syntax), and creates a list which easily
> recognizes the various patches people have contributed, or even ideas
> or bugs they have found.  The current output is here:
> http://www.red-bean.com/svnproject/contribulyzer/

Cool (and a cool name too)

Ross

Re: Part 2: Future of RAT

["Hyrum K. Wright" <hy...@mail.utexas.edu>]

On Tue, Aug 17, 2010 at 2:55 AM, Ross Gardler <rg...@apache.org> wrote:
> Something where I could say "are we overlooking a potential committer" or "do any committers appear to have gone emeritus without telling us".

You may be interested in something we've used with Subversion for
several years: the contribulyzer.  It looks at log messages (which
must have a special syntax), and creates a list which easily
recognizes the various patches people have contributed, or even ideas
or bugs they have found.  The current output is here:
http://www.red-bean.com/svnproject/contribulyzer/

-Hyrum

Re: Part 2: Future of RAT

[Ross Gardler <rg...@apache.org>]

On 17 Aug 2010, at 06:28, Stefan Bodewig <bo...@apache.org> wrote:

> On 2010-08-16, Ross Gardler wrote:
> 
>> I think the general "RAT is great" vibe of the discussion was
>> misrepresentative of what RAT currently is. There is not much to RAT
>> at present other than a very complex pattern matcher that ensures
>> license headers are present.
> 
>> It does not audit releases and to pretend it does is dangerous.
> 
> Many, many thanks Ross.  I finally start getting the whole thread.  All
> the time I knew I must be missing something but completely failed to
> grasp that people were looking at RAT for more than what it currently
> is.
> 
> I completely agree with your description of RAT's current state but I
> never expected it to be more - that's why I said its scope (not its code
> base) was too small for a TLP.
> 
>> So where is RAT going with respect to functionality. Is it really
>> "complete" in it's current form, as some people have suggested?
> 
> As one of the people who suggested it was complete, my answer is yes,
> but only because I never envisioned RAT to be more.
> 
> I just re-read <http://wiki.apache.org/incubator/RatProposal> and
> obviously RAT is not there, yet.  OTOH I doubt that extending RAT beyond
> what it currently does is on anybody's agenda, at least it is not on
> mine.
> 
> Stefan

Well, the reason I wanted to mentor the project was because I want an audit tool, and not just a release audit tool. I want tools that help with identifying and tracking communities. 

Something where I could say "are we overlooking a potential committer" or "do any committers appear to have gone emeritus without telling us". 

I'm still interested in that, the question is whether or not I'll ever find the time to actually work on it. We all know desires and wishes don't write code. 

Ross

Re: Part 2: Future of RAT

[Stefan Bodewig <bo...@apache.org>]

On 2010-08-16, Ross Gardler wrote:

> I think the general "RAT is great" vibe of the discussion was
> misrepresentative of what RAT currently is. There is not much to RAT
> at present other than a very complex pattern matcher that ensures
> license headers are present.

> It does not audit releases and to pretend it does is dangerous.

Many, many thanks Ross.  I finally start getting the whole thread.  All
the time I knew I must be missing something but completely failed to
grasp that people were looking at RAT for more than what it currently
is.

I completely agree with your description of RAT's current state but I
never expected it to be more - that's why I said its scope (not its code
base) was too small for a TLP.

> So where is RAT going with respect to functionality. Is it really
> "complete" in it's current form, as some people have suggested?

As one of the people who suggested it was complete, my answer is yes,
but only because I never envisioned RAT to be more.

I just re-read <http://wiki.apache.org/incubator/RatProposal> and
obviously RAT is not there, yet.  OTOH I doubt that extending RAT beyond
what it currently does is on anybody's agenda, at least it is not on
mine.

Stefan

Re: Part 2: Future of RAT

[Ross Gardler <rg...@apache.org>]

On 16 Aug 2010, at 21:41, Stefan Bodewig <bo...@apache.org> wrote:

> On 2010-08-16, Jochen Wiedmann wrote:
> 

...

> If RAT is going anywhere we must ensure it has enough people who care -
> of which there seem to be plenty, they may need to be recruited, though.

Well, as a mentor of RAT, a user outside the ASF and an early contributor I would most likely be involved in a small way - if there is a community to turn RAT into what it claims to be, a release audit tool. Although note I'm only a mentor and early contributor, not a committed. 

I think the general "RAT is great" vibe of the discussion was misrepresentative of what RAT currently is. There is not much to RAT at present other than a very complex pattern matcher that ensures license headers are present. 

It does not audit releases and to pretend it does is dangerous. 

So where is RAT going with respect to functionality. Is it really "complete" in it's current form, as some people have suggested?

Don't get me wrong, it's a great tool for a very limited use case (I do use it outside the ASF), but a top level project that stops at that functionality is not likely to draw more patches from me since there are other more complete tools that fill my wider audit needs (albeit not Apache licensed).

In a similar vein, If RAT really is done then why do we need contributors on board at all?

I think whether it is done or not should influence where it goes since the completeness or otherwise of the strategic goals of the project will affect whether people are on board or not. 

Ross


> 

Re: Part 2: Future of RAT

[Stefan Bodewig <bo...@apache.org>]

On 2010-08-16, Jochen Wiedmann wrote:

> now that the waves are flattened again,

The results have been interesting, looks as if many people outside of
the RAT development team feel strongly about RAT.

8-)

> Now, back to the rat-dev list, I'd like to put the same question to
> you again: Isn't it time for RAT to leave the Incubator? And if so,
> what would be the target?

In order to answer this properly we must also address Niall's concerns
that are all but unfounded.  Are all PPMC members actually aware that
they are on board?  Could everybody who is a member of the PPMC please
check they are actually subscribed to rat-private?

If RAT is going anywhere we must ensure it has enough people who care -
of which there seem to be plenty, they may need to be recruited, though.

> Regardless of what was written in the first thread, I still feel that
> the project is small in all aspects that I can think of. That's why I
> clearly thought of becoming a subproject of another, like Commons.

I understand that.  From where I stand RAT has just been a tool like
David's clutch or many other tools that have been developed to upgrade
licenses or whatever.  That's why infra or the incubator itself appeals
to me.

> But I did neither expect the pressure in the direction of becoming a
> TLP (Greg Stein being interested), nor the lack of interest from the
> side of Commons.

I do understand that Commons feels RAT doesn't fit its scope.

> But the alternative, remaining in the Incubator is something which I
> consider to be worse in the medium term.

I'd consider the Incubator as a target as a graduated subproject, not as
an eternal podling.

> Should the discussion be heading into the direction of RAT as a TLP
> (which I still view as the second best only) and the question of a
> chair become a problem, then I'd offer myself

If we get there, I for one would gratefully accept your offer ;-)

Stefan