[jira] [Commented] (FREEMARKER-97) Header capitalization using standard styles

["Daniel Dekany (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/FREEMARKER-97?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16494879#comment-16494879 ] 

Daniel Dekany commented on FREEMARKER-97:
-----------------------------------------

This is the same kind of problem than truncating long text, in that it's often needed, but has no widely accepted algorithm. Even inside the same culture there can be subtle variations. So, I think the way to go is adding `?title_case` (and `?truncate`), but make the algorithm pluggable in the configuration (on `Configurable` level, so that it can even vary per page). Furthermore the algorithm depends on the current `locale`, which must be take into account when designing this configuration API. So no wonder these "too human" transformations were avoided so far, requiring users to come up with their own functions.

Also, there need to be a default implementation. Hopefully it's possible to chose a good enough default, which's better than the 95% of ad-hoc ones users comes up with in the middle of some other work. But the sensible default is locale specific. Especially in this case, as we need to recognize some English words when the locale starts with `en`, which is quite weird (I mean, hard-coded support for a specific language). It would be interesting to know how they do this at some of the bigger nations at least, and what would be the best default if we don't explicitly support the language. (Then there's the slight problem that if we support, say, English, French and Spanish out of the box, then why not language X... people can read things into that.) Or if what exactly is the ideal rule for English, because the linked implementations differ, and I'm not sure which is the best. Any help is appreciated. Especially contributed code.

> Header capitalization using standard styles
> -------------------------------------------
>
>                 Key: FREEMARKER-97
>                 URL: https://issues.apache.org/jira/browse/FREEMARKER-97
>             Project: Apache Freemarker
>          Issue Type: Wish
>          Components: engine
>    Affects Versions: 2.3.28
>            Reporter: T X
>            Priority: Minor
>
> FreeMarker offers a couple of simple algorithms for changing the case of titles:
>  * [https://freemarker.apache.org/docs/ref_builtins_string.html#ref_builtin_cap_firs|https://freemarker.apache.org/docs/ref_builtins_string.html#ref_builtin_cap_first]
>  * [https://freemarker.apache.org/docs/ref_builtins_string.html#ref_builtin_capitalizet|https://freemarker.apache.org/docs/ref_builtins_string.html#ref_builtin_cap_first]
> Neither of these capitalize the text in ways that adhere to various standard styles:
>  * [Chicago Manual of Style|https://en.wikipedia.org/wiki/The_Chicago_Manual_of_Style]
>  * [Associated Press|https://en.wikipedia.org/wiki/AP_Stylebook]
>  * [MLA Style Manual|https://en.wikipedia.org/wiki/MLA_Style_Manual]
>  * [APA Style|https://en.wikipedia.org/wiki/APA_style]
> Consider the following texts:
>  * On iPhone the Transcript Extends Outside of Screen Frame
>  * PEAR And GNA Report Performance
>  * BCMailPlusFTPClient Sends Document without Document ID
>  * JWebUnit: Non-PEN Orders main.xhtml Meta Refresh Tag Issue
> These are correct as written and must not be adjusted by an algorithm that changes the text title. There are a couple of web sites that produce the expected titles (note that the second site capitalizes the word "without," which implies the algorithm does not use Chicago conventions):
>  * [https://titlecaseconverter.com/]
>  * [http://individed.com/code/to-title-case/]
> There are a variety of implementations that perform such a feat:
> ||Language||Link||
> |JavaScript|[https://github.com/gouch/to-title-case/blob/master/to-title-case.js]|
> |Perl|[https://gist.github.com/gruber/9f9e8650d68b13ce4d78]|
> |PHP|[https://gist.github.com/HipsterJazzbo/2532c93a18db7451b0cec529c95b53c4]|
> These implementations do not require a whitelist. (So "iPhone" and "ClassName" will remain as given.) Apache Commons' {{WordUtils}} class does not implement Chicago Style, and I suspect it is also a fairly simple algorithm.
> Since {{?capitalize}} and {{?cap_first}} are taken, I propose {{?title_case}} with an optional parameter (default is Chicago):
>  * chicago
>  * ap
>  * apa
>  * mla
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)