[Pkg-ime-devel] Bug#644361: Bug#644361: fcitx: Typos in package description

[Aron Xu]

Thanks for your help!

On Wed, Oct 5, 2011 at 19:29, Justin B Rye <jbr at edlug.org.uk> wrote:
> Martin Eberhard Schauer wrote:
>> translating the package description (1) to German the spell-checker
>> found typos.
>
> I only see one typo, but yes, it's got some problems.
>
>> Description: Free Chinese Input Toy of X
>
> Unfortunately this official expansion of FCITX makes a rather bad
> description ("Free" is redundant in Debian, "Chinese" is overspecific,
> "Toy" is misleading, and it's "for" X, not "of" X), so the long
> description has a lot of work to do.
>

Agree.

>>  Fcitx is the Free Chinese Input Toy of X, which was initially designed
>
> No need to repeat the expansion of FCITX; it might be better to say
> right at the start that
>
>   Fcitx is an Input Method engine which was initially designed [...]
>
>>  for Chinese users, and used XIM protocol. Now it has evolved into a highly
>                              ^the
>>  modularized, feature rich input method framework for Unix-like platforms
>>  supporting a considerable amount of front-ends, back-ends and modules.
>                             ~~~~~~                         ^,
> "Amount" in this sort of context (with "countable" nouns) is one of
> those things that's common in speech but frowned on by schoolteachers.
> I'd suggest s/amount/range/.
>

Agree.

>>  .
>>  It is an ideal choice for the vast majority. Many of its features make
>>  users of Unix-like platforms have a fully modern input experience for
>>  the first time. It has also greatly lowered the effort for developers,
>>  making the development of extended functions much easier than ever before.
>
> Sorry, I don't agree with your s/threshold/effort/; "lowering the
> threshold for developers" is a (rather clichéd) set expression.
> But for most of the Fcitx suite it's an irrelevance anyway - this
> description is supposed to be telling end users what they'll get when
> they install the binary package, not the sourcecode.  By all means add
> this text to the package-specific paragraph for fcitx-dev, but it
> probably doesn't belong here.
>

Agree.

> There are still some slightly clumsy phrasings here, and worse, it's
> a terrible summary of the package's advantages.  It's all very well to
> claim that it makes users have (that should be "gives" or similar) a
> good experience, but it's strange to try to *count* the unlisted
> features that do this ("many").  "Modern" in a package description is
> always a danger sign, since the features that were groundbreaking when
> it was originally written may since have become liabilities.  The
> "Unix-like platforms" part was already redundant the first time it was
> mentioned.  And all these superlatives just fill me with distrust.
> Can we have less advertising and more relevant information, please?
>

OK, though my English is poor to give a good paragraph with more
relevant information, I'd like to explain some of the facts.

Input method under Unix-like platforms falls far behind on Windows,
including input method engines as well as look and feel. Fcitx makes
some progress:

1.Much more friendly APIs for input engine developers (comparing with
IBus and SCIM).
2.More thorough modular system than IBus. As an example, it can
support Fbterm, pure Xlibs, GTK, and KDE based UIs with minimal
dependencies (they are all just plug-ins). In other word, it's a input
method framework like IBus, but with more features.
3.Much more close to the user experience of main stream input methods
on Windows, including features like skins, easy-to-use configuration
tools, and many details that was not possible for other input method
frameworks.

> For instance, if it's no longer just for Chinese (the oldstable
> package listed WuBi, Pinyin, and QuWei), what else does it support?
> And where are the docs for people who can't read the Chinese-only
> version at http://fcitx.github.com/handbook/?
>

It's ready for people to port Japanese/Korean input methods, though
still no people jumped in to write the wrapper. Upstream is willing to
provide English version of the documents, but he's also in trouble
about his English...

>>  .
>>  This meta-package pulls in a set of recommended Fcitx components, which
>>  is expected fit the needs of most desktop users.
>              ^to
>
> I (slightly) prefer "metapackage" to "meta-package", but it's a matter
> of taste.
>

OK.

> "Most desktop users" don't need an XIM in the first place, so I'd
> rephrase it slightly.
>
> So here's a work-in-progress version 3:
>
> Description: Free Chinese Input Toy of X
>  Fcitx is an Input Method engine which was initially designed for Chinese,

Fcitx isn't an engine, it's should be better called a framework. An
engine is like libsunpinyin or libgooglepinyin.

>  using the XIM protocol. Now it has evolved into a highly modularized,
>  feature rich input method framework supporting a considerable range of
>  front-ends, back-ends, and modules.
>  .
>  [ACTUAL INFORMATION ABOUT ITS FEATURES GOES HERE]
>  .
>  This metapackage pulls in a set of components recommended for most
>  desktop Fcitx users.
>

Thanks again!


-- 
Regards,
Aron Xu