On Tue, 19 Mar 2019 at 23:12, jgenender <[email protected]> wrote:
>
> Robbie Gemmell wrote
> > Saying you see value in something is not someone saying they will help
> > maintain it. There being users is not someone saying they will help
> > maintain it. I see one person who might have said they intend to help
> > maintain things on the CPP side. Some other people have either said or
> > implied they wont be helping maintain them. I've avoided commenting in
> > the thread so far as that latter direction includes me.
>
> I think a few folks here have implied or said they would help maintain it.
> This has been established.
>

At the time I sent that mail I only saw that 1 person had sort of
suggested they might after you previously said they would. I asked for
some clarity in another mail, which they gave, and a couple others
have since also chipped in.

Which is good.

>
> Robbie Gemmell wrote
> > While I am commenting though..I think the 'there are users' argument
> > goes both ways and its in general nicer to let any users know the
> > actual state of things, once established. I believe components that
> > are not under active maintenance should be noted that way in some
> > fashion so its clear thats true, in case there are actually users
> > (nothing is stopping them continuing to use it regardless).
>
> No it doesn't and thats not fair.  Case in point... I have a visibility of
> users utilizing those APIs as my work is outward facing in the ActiveMQ
> space.  I'm not an internal engineer working on specialized products.  I
> work with end users daily for multiple clients.  I am clearly aware there is
> a good solid group of people using those APIs.  I am not going to sit and
> list the user names and companies for whom I know use it.  Unless you are
> tracking the downloads, its pretty hard to say how many users leverage it.
> But I am in the know that there is a good solid base of fortune 100
> companies using those products.  So it does not go both ways.
>

I think you entirely misunderstood my point there.

I was saying that I think if there are still users of something and it
isnt actually being actively maintained, that it is good to make that
clear for them (and any potential new users). Which companies
specifically are using it for what and what type of work you do isnt a
factor to me in holding that view, mainly just who is actually
maintaining bits for users is really. That said, obviously the more
people we do know are using something, the more important I'd say it
is to be clear on its status.

>From this discussion it seems the status is there are indeed still
folks out there committing to help maintain these bits going forward,
which would mean its all good and there isn't a need to clarify the
site for users.

>
> Robbie Gemmell wrote
> > If it becomes clear thats the case, e.g it seems insuffcient folks
> > actually
> > step forward to maintain something, then to me it doesnt seem
> > particularly different to the recent commits marking Apollo as
> > deprecated, or the LevelDB related stuff being marked deprecated in
> > the past as mentioned earlier in the thread.
>
> That is an apples to oranges comparison.  Apollo was fully abandoned adn
> unanimously voted to sunset it.

After it sat going stale and unmaintained for years. It also took an
age to follow through on the vote to mark it deprecated.

A clear discussion like this around Apollo would have been great far
sooner in my view.

> LevelDB was connected to an outside library
> written by Dain Sundstom (another one of those github repos owned by a
> single person).  That LevelDB code was difficult to update without forking
> it over here.  It was unwieldy and a mess of issues.  That also was
> unanimously sunsetted.  These APIs clearly are not and its far from
> unanimous.
>

I dont recall saying anything was unanimously sunset, only that this
thread was a good means to better establish the status of the
components, and that "If it becomes clear" the status was similar,
then it would in fact be similar.

Instead a few folks have now (mostly since that mail) chipped in
saying they are looking to maintain these bits going forward, which is
great, as now we all know that and the simple discussion has in fact
clarified their status.

>
> Robbie Gemmell wrote
> > I dont think I'd consider a component maintained and releasable
> > without a site presence. Linking to old content for docs would be fine
> > though.
>
> and that is where is vehemently disagree with you.  Linux is full of
> drivers, APIs, etc, that only have a README.txt file attached to it.  If
> nobody here has time to maintain a website, then IMHO a README.md or txt is
> just fine.

I dont really see linux drivers etc as comparable personally. Download
pages, release notes, docs etc don't seem like optional things for a
component of an Apache project to me.

That said, I'm not offering to maintain these pages as I have enough
site maintainence to do elsewhere already. I am merely stating my
opinion, as you are.

> If you are adamant about a website, then pick up a shovel and
> step up to do it.  Thats an easy way to get commit.

I've had commit rights for some years now, but thanks for the tip.

> The CPP and NMS code
> bases are relatively stable and the JIRAs out there have some nice patches
> associated with them.  IMHO the main issue with those repos is getting them
> properly cleaned up for releases like how ActiveMQ does.  Jamie Goodyear has
> already discussed this on the dev lists with no response.  Yet when he
> attempted to do PRs, they got -1'd.  So its a bit of the chicken and egg on
> that one.

I recall some discussion and back and forth on dev@ a few months ago
on that and Jamie noted earlier that he sees the way forward now, so
seems like there is at least some handle on that.

>
> BTW Robbie, welcome to open source :-). That how these things go.  People
> step up to do different things and not everyone wants to do the whole
> tomato.

Indeed, which is why discussions like this are good to help bring out
who is interested in doing what. I dont see any issue or downside to
having clarifying discussions like this to help get a shared view on a
given situation, I get the feeling that some do though.

Robbie

Reply via email to