[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#513053: [checks/descriptions] warn when first person is used in descriptions



clone 513053 -1
retitle 513053 [checks/descriptions] include first-word-of-sentences-too-repetitive
tag 513053 moreinfo
thanks

Russ Allbery wrote:

> Raphael Geissert writes:
> 
>> I originally used severity: normal, but later was unsure about the
>> severity because if I filed a bug report I would use severity: minor but
>> using this severity causes an EWI code of I which lintian by default
>> doesn't display it.
> 
> These patch sets have a lot of this sort of multiple commits in them.
> Could you use git rebase --interactive before generating the mailbox to
> squash some of these commits together when they're simple corrections to
> previous diffs rather than new work?

I don't like to rebase, but might end up doing that or resetting commits if
you find it noisy to keep all history.

> 
> I'll take care of compressing this one down.
> 
>> As a bonus the attached mbox contains a patch implementing:
>>
>>> Tag: first-word-of-sentences-too-repetitive
>>> Severity: minor
>>> Certainty: wild-guess
>>> Info: The given word is used very often as the first word of a sentence.
>>> This makes the description hard to read because it is repetitive.
>>>  .
>>>  Consider rephrasing the description, using a more wide diversity of
>>>  words. In case you have troubles with it an email to the english i18n
>>>  mailing
>>> list should help.
> 
> Hm, this one fails my "would I find this helpful or just annoying if it
> triggered on one of my packages" test.  It feels very picky to me, even
> too picky for --pedantic, and getting into the area where automated
> analysis just isn't helpful compared to a native speaker reviewing the
> text.  But this is a "soft" opinion -- I'm willing to be outvoted if other
> people don't agree with me.
> 

I checked many packages and almost none triggered it. I wrote it based on a
package  descriptions review I started a week ago. I also though about using
a scores system so that "This ... This... This..." and "This... That...
This... This... That... If... That..., etc" would both trigger, but "This...
This... <two or more sentences here> This..." would not.

Further analysis of the impact of such check is needed, I admit it. I'm
cloning this report and tagging it as moreinfo to preserve the discussion of
this other check.

Cheers,
-- 
Raphael Geissert - Debian Maintainer
www.debian.org - get.debian.net





Reply to: