www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Phobos Review Queue

reply "Brian Schott" <briancschott gmail.com> writes:
std.uni was recently accepted for inclusion in Phobos, and as far 
as I'm aware there are no reviews currently in progress.

We currently have a backlog of several modules that are ready for 
comments or review[1]. There seems to be no real schedule for 
starting reviews other than "when someone pushes for it".

I'm ready to start a discussion on a D lexer module I've written 
for inclusion in Phobos. There are several modules in line ahead 
of mine, so I'm willing to hold off on this if the authors of 
those other modules are ready for review.

[1] http://wiki.dlang.org/Review_Queue
Jun 06 2013
next sibling parent "Paul D. Anderson" <nobody hotmail.com> writes:
On Thursday, 6 June 2013 at 19:50:51 UTC, Brian Schott wrote:

 std.uni was recently accepted for inclusion in Phobos, and as 
 far as I'm aware there are no reviews currently in progress.

 We currently have a backlog of several modules that are ready 
 for comments or review[1]. There seems to be no real schedule 
 for starting reviews other than "when someone pushes for it".

 I'm ready to start a discussion on a D lexer module I've 
 written for inclusion in Phobos. There are several modules in 
 line ahead of mine, so I'm willing to hold off on this if the 
 authors of those other modules are ready for review.

 [1] http://wiki.dlang.org/Review_Queue


std.decimal.bigfloat is on hold until const is fixed for big
integers. At present  const bigints can't be operands in
arithmetic functions. They can't even be copied to a mutable
bigint. (This used to work.) (In D2.060, I think.)

In the meantime, I'm re-working the code for the fixed size
decimal numbers: decimal32, decimal64 and decimal128.

Paul
Jun 06 2013
prev sibling next sibling parent reply "Brad Anderson" <eco gnuk.net> writes:
On Thursday, 6 June 2013 at 19:50:51 UTC, Brian Schott wrote:

 std.uni was recently accepted for inclusion in Phobos, and as 
 far as I'm aware there are no reviews currently in progress.

 We currently have a backlog of several modules that are ready 
 for comments or review[1]. There seems to be no real schedule 
 for starting reviews other than "when someone pushes for it".

 I'm ready to start a discussion on a D lexer module I've 
 written for inclusion in Phobos. There are several modules in 
 line ahead of mine, so I'm willing to hold off on this if the 
 authors of those other modules are ready for review.

 [1] http://wiki.dlang.org/Review_Queue


Looks like there are no objections. I'd say you're up. Just need 
a Review Manager.

Unless I'm mistaken, all that a Review Manager seems to do is 
send out the announcements of the (typically 3 week) review and 
then announce the week long voting period and finally tally and 
post the results.  I don't see why the person proposing the 
module can't be the Review Manager too.  I don't really see any 
sort of conflict of interest when the vote is done publicly and 
the vote count is usually fewer than a couple dozen.

Anyone disagree?  I think it could really help move the review 
queue along if people don't have to wrangle up a Review Manager.  
Instead they could just post like Brian has done making sure they 
aren't stepping on anyone's toes.
Jun 07 2013
parent reply "Brad Anderson" <eco gnuk.net> writes:
On Friday, 7 June 2013 at 22:41:56 UTC, Brad Anderson wrote:

 On Thursday, 6 June 2013 at 19:50:51 UTC, Brian Schott wrote:

 std.uni was recently accepted for inclusion in Phobos, and as 
 far as I'm aware there are no reviews currently in progress.

 We currently have a backlog of several modules that are ready 
 for comments or review[1]. There seems to be no real schedule 
 for starting reviews other than "when someone pushes for it".

 I'm ready to start a discussion on a D lexer module I've 
 written for inclusion in Phobos. There are several modules in 
 line ahead of mine, so I'm willing to hold off on this if the 
 authors of those other modules are ready for review.

 [1] http://wiki.dlang.org/Review_Queue


 Looks like there are no objections. I'd say you're up. Just 
 need a Review Manager.

 Unless I'm mistaken, all that a Review Manager seems to do is 
 send out the announcements of the (typically 3 week) review and 
 then announce the week long voting period and finally tally and 
 post the results.  I don't see why the person proposing the 
 module can't be the Review Manager too.  I don't really see any 
 sort of conflict of interest when the vote is done publicly and 
 the vote count is usually fewer than a couple dozen.

 Anyone disagree?  I think it could really help move the review 
 queue along if people don't have to wrangle up a Review 
 Manager.  Instead they could just post like Brian has done 
 making sure they aren't stepping on anyone's toes.


Found the description of the process: 
http://wiki.dlang.org/Review/Process

It's somewhat vague and could probably use some standard 
announcement templates people could use.

Looks like it's actually two weeks of review, one week of voting. 
  I still think the module author should be able to be Review 
Manager though.
Jun 07 2013
parent reply "Jakob Ovrum" <jakobovrum gmail.com> writes:
On Friday, 7 June 2013 at 22:52:47 UTC, Brad Anderson wrote:

 Found the description of the process: 
 http://wiki.dlang.org/Review/Process

 It's somewhat vague and could probably use some standard 
 announcement templates people could use.

 Looks like it's actually two weeks of review, one week of 
 voting.
  I still think the module author should be able to be Review 
 Manager though.


It mentions that after review, it is up to the review manager 
whether or not to continue with a vote. That seems like a big 
conflict-of-interest decision.

If I remember correctly, this has even happened before, with 
std.net.curl.
Jun 08 2013
next sibling parent reply "Jonas Drewsen" <nospam4321 hotmail.com > writes:
On Saturday, 8 June 2013 at 07:30:03 UTC, Jakob Ovrum wrote:

 On Friday, 7 June 2013 at 22:52:47 UTC, Brad Anderson wrote:

 Found the description of the process: 
 http://wiki.dlang.org/Review/Process

 It's somewhat vague and could probably use some standard 
 announcement templates people could use.

 Looks like it's actually two weeks of review, one week of 
 voting.
 I still think the module author should be able to be Review 
 Manager though.


 It mentions that after review, it is up to the review manager 
 whether or not to continue with a vote. That seems like a big 
 conflict-of-interest decision.

 If I remember correctly, this has even happened before, with 
 std.net.curl.


I am pretty sure I was not the one deciding to start a vote for 
std.net.curl back then.

Can't remember who was the review manager though.
Jun 08 2013
parent "David Nadlinger" <code klickverbot.at> writes:
On Saturday, 8 June 2013 at 10:21:58 UTC, Jonas Drewsen wrote:

 I am pretty sure I was not the one deciding to start a vote for 
 std.net.curl back then.

 Can't remember who was the review manager though.


That would have been me for the first time round, and then the 
other David (i.e. Simcha) for the second iteration after you 
incorporated the review comments.

David
Jun 08 2013
prev sibling parent reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 6/8/13 3:30 AM, Jakob Ovrum wrote:

 On Friday, 7 June 2013 at 22:52:47 UTC, Brad Anderson wrote:

 Found the description of the process:
 http://wiki.dlang.org/Review/Process

 It's somewhat vague and could probably use some standard announcement
 templates people could use.

 Looks like it's actually two weeks of review, one week of voting.
 I still think the module author should be able to be Review Manager
 though.


 It mentions that after review, it is up to the review manager whether or
 not to continue with a vote. That seems like a big conflict-of-interest
 decision.


Why? The review manager does not have a particular interest in getting 
the proposal accepted or refused; the role is there just to ensure a 
fair process.

Andrei
Jun 08 2013
next sibling parent reply "H. S. Teoh" <hsteoh quickfur.ath.cx> writes:
On Sat, Jun 08, 2013 at 09:26:09AM -0400, Andrei Alexandrescu wrote:

 On 6/8/13 3:30 AM, Jakob Ovrum wrote:

On Friday, 7 June 2013 at 22:52:47 UTC, Brad Anderson wrote:

Found the description of the process:
http://wiki.dlang.org/Review/Process

It's somewhat vague and could probably use some standard announcement
templates people could use.

Looks like it's actually two weeks of review, one week of voting.
I still think the module author should be able to be Review Manager
though.


It mentions that after review, it is up to the review manager whether or
not to continue with a vote. That seems like a big conflict-of-interest
decision.


 
 Why? The review manager does not have a particular interest in
 getting the proposal accepted or refused; the role is there just to
 ensure a fair process.


[...]

If the review manager is also the author, wouldn't he have particular
interest in getting the proposal accepted?


T

-- 
There are four kinds of lies: lies, damn lies, and statistics.
Jun 08 2013
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 6/8/13 10:19 AM, H. S. Teoh wrote:

 If the review manager is also the author, wouldn't he have particular
 interest in getting the proposal accepted?


Those should never be the same.

Andrei
Jun 08 2013
prev sibling parent "Jakob Ovrum" <jakobovrum gmail.com> writes:
On Saturday, 8 June 2013 at 13:26:06 UTC, Andrei Alexandrescu 
wrote:

 Why? The review manager does not have a particular interest in 
 getting the proposal accepted or refused; the role is there 
 just to ensure a fair process.

 Andrei


The topic is whether someone could be review manager for their 
own contribution.
Jun 08 2013
prev sibling parent reply "Jesse Phillips" <Jesse.K.Phillips+D gmail.com> writes:
On Thursday, 6 June 2013 at 19:50:51 UTC, Brian Schott wrote:

 std.uni was recently accepted for inclusion in Phobos, and as 
 far as I'm aware there are no reviews currently in progress.

 We currently have a backlog of several modules that are ready 
 for comments or review[1]. There seems to be no real schedule 
 for starting reviews other than "when someone pushes for it".

 I'm ready to start a discussion on a D lexer module I've 
 written for inclusion in Phobos. There are several modules in 
 line ahead of mine, so I'm willing to hold off on this if the 
 authors of those other modules are ready for review.

 [1] http://wiki.dlang.org/Review_Queue


I had contacted Jacob (std.serialize), but he said that he 
wouldn't be available this week so I haven't made an announcement.

As for the process, it is the Boost review with reductions to 
match our scale. For example I believe Boost specifies a "Review 
Wizard" whom manages the "Review Managers," but we need the 
reviewers first, which might be easier to do with someone to do 
some recruiting.

Brad, As for separations of concern. It probably isn't important 
with the current quantity and quality of submissions, but there 
is quite a bit of room for the review manager to work with the 
submitter to identify and resolve major problems. But yes, it is 
mostly getting the announcements out and tallying votes.
Jun 07 2013
parent Jacob Carlborg <doob me.com> writes:
On 2013-06-08 02:11, Jesse Phillips wrote:


 I had contacted Jacob (std.serialize), but he said that he wouldn't be
 available this week so I haven't made an announcement.


I'm available now.

-- 
/Jacob Carlborg
Jun 09 2013