[review queue] What to do about the library review queue?

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
158 messages Options
1 ... 5678
Reply | Threaded
Open this post in threaded view
|

Re: Attn: New Boost library policy text ready for approval

Boost - Dev mailing list
On 3/20/17 2:56 PM, Niall Douglas via Boost wrote:

>>>> Back on topic, I think that the current process of getting a library
>>>> into the review queue is a bit outdated. I suggest we make use of >
>>> existing infrastructure and make a Github repository "review" owned by
>>>> the Review Wizard in which submissions occur by way of the endorsing
>>>> Boost member creating an issue with the description of the library.
>>>
>>> The problem with this is that we don't want one member endorsing a
>>> library for review. We want *lots*.
>>>
>>> Almost without doubt when potential review managers scan the list of
>>> review pending libraries, they will prioritise those libraries with
>>> the most public endorsements.
>>
>> How would they know which libraries have the most endorsements?
>
> I had been thinking that an extra column in the table at
> http://www.boost.org/community/review_schedule.html would be entitled
> "Seconded by" and in the cell would be the names of all those who
> endorsed that library for review. In other words, if you Peter ask for
> endorsements for review here for your new library SharedPtr2 or
> something, and say Edward, myself, Robert, Beman and Michael all
> publicly say "this library looks very likely to pass a review", all our
> names appear in that column.
>
> Does this make sense now?

LOL - not to me.  If I really studied it it might.  But right now I'm in
TMP/constexpr hell and my brain is already overloaded.

>
> I would not recommend a "click to star" type system for endorsing a
> library for review. Endorsing a library for review publicly with your
> name is a solemn statement that you have given a cursory check of the
> library and that you publicly declare you think it has some chance of
> passing a review.
>
> I have no opposition to a *separate* "click to star" upvoting system so
> people can easily upvote some review as being more urgent than others.
>
> Niall

The whole thing seems pretty complex to me.  I'm not sure who is going
to administer/script such a thing.  FYI the incubator has the facility
for interested parties to submit a review along with star ratings.  The
idea was that a library would get officially considered only once some
number of reviews indicated interest.  Problem is/was - very few reviews
were submitted.  So I'm a little skeptical that anything more elaborate
than that is going to be successful.

Robert Ramey


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: Attn: New Boost library policy text ready for approval

Boost - Dev mailing list
In reply to this post by Boost - Dev mailing list
> The new procedure would be, the library author posts to the list, the
> first Boost member who endorses the library creates an issue for the
> library, additional endorsements are posted as comments on that issue.
> The Review Wizard, being the owner of the repo, receives notifications
> for each and can respond in the issue, if needed.
>
> That is, all the communication that is relevant for a particular
> submission reaches the Review Wizard without him having to read the
> Boost list, or relying on people notifying him via private e-mail.
>
> One exception to all the communication taking place in the issue could
> be when a review manager is suggested, because in case of rejection,
> people may not appreciate this taking place in public. I don't know how
> often review managers are rejected though, so I don't know if this is a
> practical problem.
>
> Re Github stars, if the library is already on Guthub, which it pretty
> much has to be, the repo stars can be used as a metric of popularity.
> (The other way to gauge popularity is by the number of comments the
> library submission issue has attracted.)

I think this is all far too much for people to hold in their heads.

If you were to script all this up into some webforms which called the
github api for people, I could see it working.

But for people to remember to open tickets and +1 them and all that ...
I don't think it will fly. Too much to remember.

(and Robert basically says the same thing in another reply)

Don't get me wrong for a second Peter, I would *just love* to automate
the hell out of all this stuff, bring state of the art automation to
bear on "the Boost scalability problem". But that stuff doesn't come
cheap, neither to implement, debug nor to maintain. Plus, nobody here
seems keen on automation, we seem to like static HTML and email, and
anything exceeding that is a very tough sell, especially anything
regarding changes to process.

Niall

--
ned Productions Limited Consulting
http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: Attn: New Boost library policy text ready for approval

Boost - Dev mailing list
Niall Douglas wrote:

> I think this is all far too much for people to hold in their heads.
>
> If you were to script all this up into some webforms which called the
> github api for people, I could see it working.

There's no difference between filling out some web form that calls the
Github API, or filling out the Github web form that creates an issue.

> But for people to remember to open tickets and +1 them and all that ... I
> don't think it will fly. Too much to remember.

Who are those people who have to remember? Whoever endorses the library
creates the issue (or adds his recommendation to it). This isn't rocket
science.


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: Attn: New Boost library policy text ready for approval

Boost - Dev mailing list
In reply to this post by Boost - Dev mailing list
> The new procedure would be, the library author posts to the list, the
> first Boost member who endorses the library creates an issue for the
> library, additional endorsements are posted as comments on that issue. The
> Review Wizard, being the owner of the repo, receives notifications for
> each and can respond in the issue, if needed.
>
> That is, all the communication that is relevant for a particular
> submission reaches the Review Wizard without him having to read the Boost
> list, or relying on people notifying him via private e-mail.

It occurs to me that this would also work for the actual reviews of the
library. Those could be posted as comments on the issue as well. (Or in a
"Review: libname" issue distinct from the "Submission: libname" one, to
avoid clutter.)

This would be much easier on the review manager who currently needs to scan
the high-volume Boost list in order to collect the reviews. With the issue
system, all reviews would already have been collected in one convenient
place.


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: Attn: New Boost library policy text ready for approval (was: Re: [review queue] Proposed new policy to enter the review queue)

Boost - Dev mailing list
In reply to this post by Boost - Dev mailing list
On Sun, Mar 19, 2017 at 1:30 PM, Niall Douglas via Boost <
[hidden email]> wrote:

> On 17/03/2017 12:51, Niall Douglas via Boost wrote:
> > Ok, I've asked boost-steering for feedback on this policy change. If
> > they approve, I'll do up a beta of the Boost website for people to
> > check, and if all okay it'll go live.
> >
> > The boost-steering policy change discussion request can be found at
> > https://groups.google.com/d/msg/boost-steering/rJPWdYodmtQ/JjaS-Kj4BgAJ
> > for those interested.
>
> There has been no objection to the proposed new policy for entering the
> review queue from boost-steering.
>
> The proposed reformed policy page for submitting a library for review to
> Boost can be found at
> https://boost-website.nedprod.com/development/submissions.html. If you
> object to this new policy page, now is the time to say.
>
> I'll give it until end of Wednesday 22nd March before I issue the pull
> request to boostorg/website.
>
> If there is no objection, once the new policy is onto the public
> website, we'll clear from the review queue all libraries without review
> managers attached to them. If you had been thinking of volunteering to
> review manage a library in the queue, now is the time to make yourself
> known.
>

Thanks Niall for putting this together! A few comments:


   - "Otherwise, you will just end up wasting everyone's time." This
   sentence is a bit abrasive. I suggest striking it.
   - "If what you really want is a site that will just post your library
   without even looking at it, you should go elsewhere." likewise
   - "and the emotional demands of a formal review" likewise. If you want
   to give the reader a heads up about the ego-crushing force of critiques by
   brilliant people, I would probably word it differently.
   - "Too often" . . . this sounds a bit rantish.

Actually, I think I'll stop here for now. I like the idea of this and I
think you've got a lot of good ideas in here. The wording needs a little
TLC IMHO. Would you be willing to let me take a stab at massaging it a bit?

-- David

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: Attn: New Boost library policy text ready for approval (was: Re: [review queue] Proposed new policy to enter the review queue)

Boost - Dev mailing list
David Sankel wrote:

> Thanks Niall for putting this together! A few comments:
>
> - "Otherwise, you will just end up wasting everyone's time." This sentence
> is a bit abrasive. I suggest striking it.
> - "If what you really want is a site that will just post your library
> without even looking at it, you should go elsewhere." likewise
> - "and the emotional demands of a formal review" likewise. If you want to
> give the reader a heads up about the ego-crushing force of critiques by
> brilliant people, I would probably word it differently.
> - "Too often" . . . this sounds a bit rantish.

It should be noted that these are not Niall's words. They come from the
original version of the page.

(It is for this reason that it would have been better to review a diff.)


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: Attn: New Boost library policy text ready for approval

Boost - Dev mailing list
In reply to this post by Boost - Dev mailing list
> Thanks Niall for putting this together! A few comments:
>
>   * "Otherwise, you will just end up wasting everyone's time." This
>     sentence is a bit abrasive. I suggest striking it.
>   * "If what you really want is a site that will just post your library
>     without even looking at it, you should go elsewhere." likewise
>   * "and the emotional demands of a formal review" likewise. If you want
>     to give the reader a heads up about the ego-crushing force of
>     critiques by brilliant people, I would probably word it differently.

These three are from the original page before this round of editing.
It's possible I wrote them once, I'd have to check the history as I've
changed that page quite a few times. But I also may not have done.

You make a valid point though. One of my aims was to make the page more
inclusive and welcoming and encouraging. The above could be better worded.

>   * "Too often" . . . this sounds a bit rantish.

Good catch.

> Actually, I think I'll stop here for now. I like the idea of this and I
> think you've got a lot of good ideas in here. The wording needs a little
> TLC IMHO. Would you be willing to let me take a stab at massaging it a bit?

If you're volunteering, then rock on, I would find that very helpful.

I've pushed the current draft to my fork of boostorg/website and you can
find it at
https://github.com/ned14/website/commit/50a9b67aa14dba610a81c1b7542a3bbf8abb97a8.
This also provides a diff for those interested in that. I've invited you
@camio to collaborate on that repo, it should give you commit and push
access.

Once you've made your changes I can pull them onto
boost-website.nedprod.com so we can all see them here for review.

Niall

--
ned Productions Limited Consulting
http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [review queue] What to do about the library review queue?

Boost - Dev mailing list
In reply to this post by Boost - Dev mailing list
> Once Boost libraries are decoupled and autonomous, the "Boost branding"
> would take on a different meaning, as users could more freely and
> deliberately pick what Boost library they want as part of their
> development (as well as runtime) platform.

This is something that certainly confused me when I first came to boost
- initially I read that boost libraries were meant to be as independent
as possible in the author docs, but then the actual libraries tend to be
highly interdependent.
Which creates exponential complexity in terms of maintenance.

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [partly OT] Re: [review queue] What to do about the library review queue?

Boost - Dev mailing list
In reply to this post by Boost - Dev mailing list
Le 17/03/2017 à 10:18, Oswin Krause a écrit :
>
>> Could you explain what is the problem having to use also Boost.Chrono?
> Let me answer your question with another question: What good is it to
> force boost.chrono on the user? If my application does not profit from
> using boost::chrono over std::chrono why should i have to use it?
>
Well, don't use it then. Who is forcing you to use Boost.Chrono?

> Here is how i see it:
> std::chrono
> + fully described by the standard and supported by all major c++
> environments
> + described in all online standard references for c++
> + we can soon expect all newcomers to know about std::chrono
> + can be used without adding another dependency on an external library
>
> boost::chrono
> + larger functionality
> + very similar, but not the same as std::chrono
A report on any difference is welcome. I'm not saying that there are not.
>
>
> mixing both:
> + converting between time points?
I've not think about this, but a patch is welcome.
> + differences in time measured /different clocks?
Chrono library doesn't specify how to convert between clocks. Neither
std nor boost. Could you clarify your concern?
> + why should there be two very similar libraries mixed in the same code?
>
>
I suspect that because you have some parts of your code that use
Boost.Chrono because std::chrono is not available.However you have other
parts that can use std::chrono. In this case, I would suspect that for
the parts that interact you will use Boost.Chrono.

Best,

Vicente


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [partly OT] Re: [review queue] What to do about the library review queue?

Boost - Dev mailing list
Vicente J. Botet Escriba wrote:

> Well, don't use it then. Who is forcing you to use Boost.Chrono?

Boost.Thread, as stated in the original message?

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [partly OT] Re: [review queue] What to do about the library review queue?

Boost - Dev mailing list
On 3/23/2017 11:56 AM, Peter Dimov via Boost wrote:
> Vicente J. Botet Escriba wrote:
>
>> Well, don't use it then. Who is forcing you to use Boost.Chrono?
>
> Boost.Thread, as stated in the original message?

This is why I wrote CXXD for other developers to use. Inevitably I
foresaw that someone would say "Why does your library use boost XXX when
I want to use C++ standard XXX instead ( or vice versa )." And yes, CXXD
supports chrono as one of its dual libraries.

This is off-subject for this post but I am glad for the opportunity to
prove a point regarding CXXD.


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: Attn: New Boost library policy text ready for approval

Boost - Dev mailing list
In reply to this post by Boost - Dev mailing list
On Tue, Mar 21, 2017 at 7:06 PM, Niall Douglas via Boost <
[hidden email]> wrote:

> > Thanks Niall for putting this together! A few comments:
> >
> >   * "Otherwise, you will just end up wasting everyone's time." This
> >     sentence is a bit abrasive. I suggest striking it.
> >   * "If what you really want is a site that will just post your library
> >     without even looking at it, you should go elsewhere." likewise
> >   * "and the emotional demands of a formal review" likewise. If you want
> >     to give the reader a heads up about the ego-crushing force of
> >     critiques by brilliant people, I would probably word it differently.
>
> These three are from the original page before this round of editing.
> It's possible I wrote them once, I'd have to check the history as I've
> changed that page quite a few times. But I also may not have done.
>
> You make a valid point though. One of my aims was to make the page more
> inclusive and welcoming and encouraging. The above could be better worded.
>
> >   * "Too often" . . . this sounds a bit rantish.
>
> Good catch.
>
> > Actually, I think I'll stop here for now. I like the idea of this and I
> > think you've got a lot of good ideas in here. The wording needs a little
> > TLC IMHO. Would you be willing to let me take a stab at massaging it a
> bit?
>
> If you're volunteering, then rock on, I would find that very helpful.
>
> I've pushed the current draft to my fork of boostorg/website and you can
> find it at
> https://github.com/ned14/website/commit/50a9b67aa14dba610a81
> c1b7542a3bbf8abb97a8.
> This also provides a diff for those interested in that. I've invited you
> @camio to collaborate on that repo, it should give you commit and push
> access.
>
> Once you've made your changes I can pull them onto
> boost-website.nedprod.com so we can all see them here for review.


Thanks. It looks like you already made some tone adjustments. Much better.
I did some more wordsmithing in https://github.com/ned14/website/commit/
72e3e7ce4351282a2da6d031633e7b0268cb36bc

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [review queue] What to do about the library review queue?

Boost - Dev mailing list
In reply to this post by Boost - Dev mailing list
On 16/03/2017 19:49, Joaquín M López Muñoz via Boost wrote:

>
>> Em 15 mar 2017, às 21:05, Ion Gaztañaga via Boost <[hidden email]> escreveu:
>>
>>> On 15/03/2017 9:22, Andrzej Krzemienski via Boost wrote:
>>> 2017-03-14 13:01 GMT+01:00 Niall Douglas via Boost <[hidden email]>:
>>>
>>> Actually, I have a question to Joaquín Mª López Muñoz and Ion Gaztañaga.
>>> What does it mean that the library in the queue has a review manager, but
>>> does not have any time slot scheduled?
>>
>> Just that Joaquín and I didn't agree on a date. But it could start before May.
>
> Basically any slot in April works fine for me.

It seems that April is full in the review schedule. Ronald & John, can
we schedule PolyCollection just after SIMD?

Best,

Ion

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [review queue] What to do about the library review queue?

Boost - Dev mailing list
In reply to this post by Boost - Dev mailing list
On 3/15/17 1:22 AM, Andrzej Krzemienski via Boost wrote:

> Another thought. When I look at the review queue, and also at the libraries
> listed in BLIncubator, my personal feeling is that some libraries do not
> fit into Boost. This is just my impression, but it rises a question.
>
> There is no bar for libraries to be requested for a formal review, without
> a review manager. Also, authors for some libraries maybe just want to get
> some useful feedback, and not necessarily get their library into Boost.
> Maybe we need some additional stage. BLIncubator was designed to fill this
> gap. Maybe it can still be made to work. Maybe people who feel something
> need to be done in the review queue, should go through the list of
> libraries in BLIncubator, and give their authors feedback.
>
> Maybe, we should be doing some informal pre-reviews. Take one library from
> the queue. Contact the author; check if he/she is still alive, and discuss
> with him why they want the library into boost and why we don't (or do) like
> it, and what we would rather expect.

Right.  It would be very easy to make a rule that no library can go into
the review queue until it has two reviews in the incubator.

I think this would address ALL the problems mentioned in this thread
with zero additional overhead, rules or bureaucracy.  It would be easy
to modify or retract as well.  In fact I think the review wizard could
easily say (if he wanted to).  Due to the high demand and limited
resources for quality boost reviews,

"I'm announcing that I won't be adding any reviews to the queue that
don't have at least two reviews already and someone willing to act as
review manager."

What would be the matter with this?

Robert Ramey

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [review queue] What to do about the library review queue?

Boost - Dev mailing list
In reply to this post by Boost - Dev mailing list
Hi Ion,

Does May 1 - 10 work for you?

Best,
Ron


> On Mar 24, 2017, at 2:41 PM, Ion Gaztañaga <[hidden email]> wrote:
>
> On 16/03/2017 19:49, Joaquín M López Muñoz via Boost wrote:
>>
>>> Em 15 mar 2017, às 21:05, Ion Gaztañaga via Boost <[hidden email]> escreveu:
>>>
>>>> On 15/03/2017 9:22, Andrzej Krzemienski via Boost wrote:
>>>> 2017-03-14 13:01 GMT+01:00 Niall Douglas via Boost <[hidden email]>:
>>>>
>>>> Actually, I have a question to Joaquín Mª López Muñoz and Ion Gaztañaga.
>>>> What does it mean that the library in the queue has a review manager, but
>>>> does not have any time slot scheduled?
>>>
>>> Just that Joaquín and I didn't agree on a date. But it could start before May.
>>
>> Basically any slot in April works fine for me.
>
> It seems that April is full in the review schedule. Ronald & John, can we schedule PolyCollection just after SIMD?
>
> Best,
>
> Ion


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [review queue] What to do about the library review queue?

Boost - Dev mailing list
On 28/03/2017 5:28, Ronald Garcia via Boost wrote:
> Hi Ion,
>
> Does May 1 - 10 work for you?
>
> Best,
> Ron

Due to local holidays, May 3-12 would be better for Joaquín and me.

Best,

Ion

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [review queue] What to do about the library review queue?

Boost - Dev mailing list
Thanks for letting me know.  I have scheduled the review.

Best,
Ron

> On Mar 28, 2017, at 5:21 AM, Ion Gaztañaga <[hidden email]> wrote:
>
> On 28/03/2017 5:28, Ronald Garcia via Boost wrote:
>> Hi Ion,
>>
>> Does May 1 - 10 work for you?
>>
>> Best,
>> Ron
>
> Due to local holidays, May 3-12 would be better for Joaquín and me.
>
> Best,
>
> Ion


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [review queue] What to do about the library review queue?

Boost - Dev mailing list
On 28/03/2017 18:25, Ronald Garcia via Boost wrote:
> Thanks for letting me know.  I have scheduled the review.
>
> Best,
> Ron

Thanks!

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
1 ... 5678