John McCabe-Dansted wrote:
> > But is difficult, since the web page does not supply any language data.
>
> I'm just noting that I have never personally found a case where the
> automatic addition of language tags is useful.
Actually, I would find it useful, if it would be possible technically ;-)
(who knows what semantic web will bring us).
> >> Admittedly, Leaving it as UK might be useful if it was a direct quote,
> >> and had proper quotation marks around it. (And yes I realize that
> >> detecting quotations in every possible language could be difficult,
> >> even for languages that actually have quotation marks. However I use
> >> citations rather than quotes, so again this isn't something that makes
> >> a difference)
> >
> > I don't see why quotes make a difference, in general (although I see
> > people usually mark quotes more often). In any case, trying to fiddle
> > with this is bound to fail.
>
> Some programs have "paste as quotation" in addition to plain "paste"
>
> However mixing UK and US English is almost never correct. I don't
> think you've mentioned any case where mixing UK and US English is
> correct. I am mostly trying to understand why you think that the very
> common requirement that a document be formatted consistently with
> either US or UK English but not both is fundamentally incorrect.
I would consider it appropriate if I would use a term of a UK english book. If
I would refer to, say, a term from linguistic book of Cambridge university
press, e.g. "contextualisation", there would be cases where I would keep the
UK spelling (for instance, if this concept is bound to a specific, British
author.
And if not, I would change the spelling and reset the language ;-)
> Maybe you believe this requirement is incorrect because the following
> should be considered acceptable?
> a) I went to Colour Press and got some color prints done.
> b) I think that Blue is a nice colour, but Bob wrote "Blue is an ugly
> color".
>
> Out of
> 1) Blue a nice colour, but green is an ugly colour.
> 2) Blue a nice color, but green is an ugly color.
> 3) Blue a nice colour, but green is an ugly color.
>
> I consider only 1 and 2 correct.
Me too.
> I don't consider 3 to be correct,
> anymore than I consider the eletelephony* poem to be proper English. I
> do not consider
> 4) Blue a nice colour, but green is an ugly
> \foreignlanguage{american}{color}. to be correct either. At best we've
> replaced a spelling mistake with a different form of mistake.
>
> Would you consider it correct if my Grammar Checker flagged/removed
> the \foreignlanguage{american} in (4) above, leaving only the forms in
> (a) and (b)?
I would expect a good grammar checker to warn me:
"Spelling not concise: you use two different spelling variants here. Please
consider revising."
I would also welcome such a feature in case of two different (but allowed)
spelling variants of a given language, say "Typographie" and "Typografie" in
German.
But I would not like it if the Grammar checker would change that without
asking me.
So maybe a good Grammar checker would consider the use of different varieties
of a language (such as UK and US English). Language markup could actually help
such a Grammar detection routine.
> Another possible solution for this annoyance: if LyX-GrammarChecker
> was improved and integrated to the point where it could quickly and
> easily remove inappropriate language tags in English texts. Currently
> LyX-GC is an unashamedly English only tool, but it could easily be
> extended to any language supported by lanuagetool (or by perl regex's
> for that matter). In any case, a LyX-GC cannot be expected to be
> lanuage agnostic; so I can see no objections to cramming English
> specific rules on quoting into LyX-GC.
I would welcome very much an improved LyX-GC.
> >> >> In principle I may be submitting a document to an organization that
> >> >> requires that all text be in Language X (and only language X), in
> >> >> which case any LyX document I submit that contains language markup is
> >> >> wrong, just as if I had included a Chapter in an article.
> >> >
> >> > But not if you use different languages (if this organization is not
> >> > completely crazy, that is).
> >>
> >> If I use a Chapter is a journal article, then it clearly should be
> >> formatted as a Chapter, rather than e.g. standard?
> >
> > I'm talking about language markup, not chapter markup.
>
> I don't see what the difference is. In the case of pasting a chapter
> into an article LyX removes the chapter because it would be incorrect
> in an article. The result isn't "correct", you can't turn a book into
> a proper article by removing all chapter tags, but at least it puts
> into the right format. LyX leaves it up to the user to fix this up.
The difference is: an book chapter is no book chapter anymore if pasted to an
article. A Frenc text is still a French text if pasted to an English text.
> In the case of pasting UK text into a US document, LyX actively adds
> the foreignlanguage tags. These tags are almost never appropriate. The
> user manually removes the tags, then fixes any spelling mistakes.
No. The tages are also there in the UK English document. They are just not
visible. If everything is UK English, there's no need to mark it. Still in
both cases the text has the language attribute UK English.
> I don't see why the following is any more incorrect than LyX's current
> chapter example: The user pastes some text into a journal format where
> only UK English articles are accepted. LyX automatically strips all
> the invalid \foreignlanguage{american} tags. The text may not be
> correct (it may well be). Again the user is responsible for fixing
> this up, a simple spell check suffices.
My point is that you lose valuable information (and hints) if you remove the
markup, but not change the actual language.
> I am not sure why it would not be correct. If we hardcode English (UK)
> and write English (UK) text, then the output will be correct?
This I do not understand.
> > Dunno about French. However, when writing German, I welcome very much
> > that the information about texts being old German spelling, new German
> > spelling or Swiss German is not lost during copying.
>
> I don't see how this would assist you. Is it normal in German to mix
> spellings in a single document, or does the meaning of a word depend
> on the dialect, or ...?
It is common to keep the original spelling. We had a spelling reform in 1998,
and still some authors use the old spelling. If I quote texts from such books,
I am supposed to keep the old spelling for that quote. Markup is crucial here,
since the hyphenation rules changed significantly, so LaTeX must hyphenate
such quotations according to the old spelling rules.
The same holds for quoting from Swiss texts, where you would usually write
"Strasse" instead of "Straße", which again implies specific hyphenation rules.
> AFAICT, In English, at least, it is almost never considered correct to
> mix sub-languages, while most US and UK words have the same spelling.
I'm not sure, at least when it comes to quotations or the use of terms.
> Also if people don't even spellcheck (or proofread) the document
> before submitting it they may as well use Wordpad or IRC. I don't see
> any point aiming LyX at this use case.
Of course they are supposed to spellcheck anyway.
> Since LyX doesn't have an English (Australian) option presumably all
> my documents are incorrect anyway. Might as well be consistently
> incorrect.
Well, this is a missing babel feature.
> >> Also if this dialog popups up immediately when you attempt to paste
> >> the text in, it would be obvious to you which source document had the
> >> wrong (sub) language.
> >
> > Such a dialog would be much more annoying than the language markup.
>
> LyX 1.6.2 gives me 5 warnings that various LyX files have a different
> textclass to lyxmacros.lyx each time I do a "View PDF" on my paper.
This is annoying, and I have fixed that for LyX 2.0 (now you can just click
"Do not warn me again" and be done).
> I think I can handle a dialog that occurs at most once in a documents
> lifetime :)
Me not, if the dialog can be ommitted.
Jürgen
