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

Re: [SPAM] - Re: [RFR] Description of new memchan package - Bayesian Filter detected spam



On 5/30/07, Justin B Rye <jbr@edlug.org.uk> wrote:
Sergei Golovan wrote:
> I'm preparing new memchan package for uploading to Debian distribution
> and asking someone to review its description.

Okay.  Diff and revised version of the text attached.

Thank you very much for so detailed review.


> Package: memchan
> Section: libs

I should start by saying that I'm not a Tcl programmer; in fact I
would have guessed that a library package for Tcl would be given a
name like libmemchan-tcl...

I'm still in doubt about the package name because all Tclers know this
package as memchan (even Memchan), but since apt-cache and aptitude
allow searching for substrings libmemchan-tcl is better than simple
memchan.


> Description: Tcl extension, which implements in-memory channels (runtime)

The comma is wrong here (before a "restrictive" relative clause).

Distinguishing between memchan and memchan-dev by labelling the
former as "runtime" and the latter as "development files" seems
unclear to me... But then again I rarely use the word "runtime", and
when I do, I mean a virtual environment, not the sort of library
that needs to be installed as a normal dependency package.  Maybe:

  Description: Tcl extension for in-memory channels - runtime library

This looks better for me.


What do better-informed people think?

> Allows to create I/O channels,

You can't "allow to do" - you can only "allow somebody to do", or
(the solution here) "allow something":

  Allows the creation of I/O channels,

> which store the data placed into them
> in memory, not on disk.

I'm not sure you need to say "the data placed into them"; it might
be enough to say "which store data in memory, not on disk".

Yes, the former was too wordy.


>                         Several channel types are implemented: fifo,
> null, random and zero channels. Also, C API is provided for creating
> custom memory channels.

Missing article ("A C API"); but let's switch the whole thing out of
passive voice:

  It implements the channel types fifo, null, random and zero; and
  provides a C API for creating custom memory channels.

> This package contains runtime library to be used in Tcl programs.

Article again, probably definite: "the runtime library".

> Homepage: http://memchan.sourceforge.net/

This pseudoheader is conventionally formatted with one extra leading
space:

   Homepage: http://memchan.sourceforge.net/

(See Policy 5.6.13 on the effects.)

I see.


> Package: memchan-dev
...
> This package contains files to be used for development of C-based
> extensions,
> which use memchan library.

Inappropriate comma, and at least one missing article:

It's my usual mistake (we don't have articles in Russian).


  This package contains files to be used for development of C-based
  extensions which use the memchan library.

Assuming this still means what it's supposed to.

Yes. The meaning is exactly what I want it to be. Thanks again.

--
Sergei Golovan



Reply to: