- From: Richard Ishida <ishida@w3.org>
- Date: Mon, 12 Jan 2004 19:39:08 -0000
- To: <public-i18n-geo@w3.org>
I've been making edits. Notes below, if you're interested... > Discussed the Apache language negotiation faq. > Here are the comments I remember, since the detailed notes I > took were promptly, accidentally vaporized after the call.... > Please check that this reflects your comments completely and > accurately. > Apologies in advance. > > > 1) The question should not reference content authors, as they > may not be the ones to address negotiation. Administrators > can also take advantage of these commands. (I do note in para > 3 and 4 the faq presumes the author, and distinguishes privs > coming from the admin.) Changed. > > 2) The faq mentions multiple approaches but then proceeds to > discuss only one (multiviews). It was suggested that we > clarify that within the faq and also in the question. Wording > of the Q (to resolve pts 1 and 2), could be something like: > > How does one set up language negotiation via Multiviews on an > Apache Web Server? Changed. It's a little long now, but avoids assumptions about understanding 'language negotiation' and adds reference to multiviews. Had I made the call, the odd one-sidedness was one of the key things I wanted to discuss - I think it works much better now. Various changes made to the text to accommodate the focus on multiviews only. > > 3) Answer, para 1 > We didn't think that the lang. neg. should be presented as an > either/or choice, since in the previous meetings on this > topic we considered it was likely people would do a combined approach. It isn't. It says 'in some cases'. > > 4) The term "language negotiation" should be defined. Some > readers may not know what it is, and even though it is a > standard term in the industry some of the participants didn't > think it was apt. It isn't really a negotiation since a > request is made, and some response is given and end of story... Done, in the background. > > 5) para 2, bullets 1 & 2 > The term "strategy" didn't seem appropriate. In bullet 1, > consider "convention" instead. In bullet 2, "strategy" also > indicated more policy-making than is required, when it is > just identifying default values or responses. > Consider something more along the lines of "planning > languages to be returned, when languages are requested that > you don't support". > > 6) para 3 > delete the word "unfortunately". It's ambiguous as to what is > unfortunate, and it is not needed. Done. > > 7) para 4 > Split the second sentence into 2 sentences. > The para should also say something to the effect: "This faq > describes only the Multiviews approach." Split using ; > > 8) The heading "One possible approach" should be "Using > Multiviews to configure Language Negotiation", "The Multviews > Approach", or some such. Removed. > > 9) We suggest the default be something other than english. See below. > > 10) Section on File Naming > Should other file types be considered besides html? (xml, css, jpg...) Referred to later. Want to keep this simple. > > 11) para 3, > 1st sentence refers to accessing the files when not on an > apache server. > Perhaps it should say when accessed directly, or when not > accessing thru the network to the server. After all, even > while being served, they can be accessed directly locally. > Similarly, the next phrase should refer to "accessing the > resource via the apache server." > > 12) Just curious, given the example, the multiviews must > always refer to html. If there were 2 files example.jpg and > example.html, the href without the extension would return the html? See below. > > 13) reference to rfc 3066 should also be in the links section. Done. The references hadn't been developped yet. > > 14) More generally, there should be many more links inline in > the text, so people can reference immediately if they choose, > and not have to look to the links section to see if there is > a link there. I hadn't worked out all the links yet. I added some. > > 15) The name of the file .htaccess is not optional is it? If > it is mandatory, the sentence should not say "typically be called". It is optional. See below. > > 16) Not discussed in the meeting, but I think the sentence > recommending a default of English should be deleted. The > widespread nature of english is irrelevant unless your page > is directed for the public at large. The choice of default > should reflect the makeup of your target audience and might > very well be some other language. And english can't be a > default if you haven't translated to english. It's just an example. But typically English is a good choice for a lingua franca. > > 17) specifying the default files 2.030 and above > ForceLanguagePriority does not specify the default, the > prioritization of languages does. FLP suggests the > methodology. Also, LanguagePriority is not specifying a > particular file, but the language. The first sentence should > be rewritten. Done. Both are needed. > > 18) Some text should explain the logic behind the example, as > opposed to stating the result. > FLP allows a fallback method to be used, and the Lang.Pri. > declares the ordering of possible choices. If a Spanish page > doesnt exist, and if an English page does exist, then it will > be served. If not, then a French page will be served. and so forth. > (This is what we speculated the behavior to be. It needs > confirmation.) I'm trying not to make this the length of a chapter in a book. I added text to say follow the links for the detail. > > 19) section on default files in 1.3.4 and above > The heading should instead refer to versions 2.0.30 and > below, or if we think there might be someone still using a > version less than 1.3.4, than "versions 1.3.4 to 2.0.30". Done. > > 20) aha! I just read the BTW. Suggest mentioning that files > other than html can be used, within the answer. With a > section name like "BTW", idiots like me may not read it, as > it sounds extraneous. However, the point about extensions is > fundamental. Done. > > 21) Background > First sentence would be slightly less ambiguous if it added > "user's" to " language preferences". Also, perhaps use active > voice to improve readability. e.g. > > User agents provide information about the user's language > preferences in the HTTP Accept-Language header of the request > to a server for a document. > (Also provide a link to accept-language header) Changed. > > 22) Localisation, practise and other terms use British spelllings. > We were hoping Santa brought you an American spell-checker! ;-) He didn't need to. I have one, but hadn't yet used it. Btw, there was ONLY those two errors, and the spell-checker doesn't pick up on the second! ;) > > 23) There is a question of how to use the recommendation of .pl-pl. > Is it: > addlanguage pl .pl-pl > LanguagePriority en pl fr > > i.e. pl-pl is the extension, it gets assigned a language > code, and then LanguagePriority uses language codes? > > Perhaps use polish throughout to make it clear. > > 24) The link for rfc 1766 is wrong. It needed replacing anyway. > > 25) Does the order of the statements in the .htaccess file > make a difference? if so what is the preferred ordering? If > not, state so. No. > > 26) Does the .htaccess only apply to the current directory, > or are subdirectories also affected? (I think subdirs are > affected, unless they have their own .htaccess). Subdirectories are affected. > > > 27) The group felt that with the changes we recommended the > faq could/should be published. > > Thankyou.
Received on Monday, 12 January 2004 14:39:09 UTC