I think, Unicode says (please correct my if I am wrong), for example, Bangla can only use code points U+0980 to U+09FF. And also, codepoint X will represent letter Y.
Who specify which letter will be placed on what location in the keyboard? Is it also the job of unicode?
For example, Who specify which letter will be placed on what location in keyboard for Bangla? I am asking this because there seems to be a dozen keyboard layouts for bangla unicode Keyboards.
What is the deal here?
There is no unified standard or single standardization body for this. The individual layouts might not be standardized at all, e.g. developed by individual companies for their products, or be standardized by some national standardization body, governmental or non-governmental, for any give language/script.
This can also be seen in the Wikipedia list of some common Bengali keyboard layouts here that you are referring to in your question as well. For each layout the respective organization that developed/standardized them is also given. (Note that the site you linked is not actually Wikipedia, but a third-party interface to it.)
Related
My organisation now has a policy to make all documents accessible, and a long list of guidance how to do so. Most of the requirements are OK. Alt-text is only a small amount of work; sans-serif fonts may not be as nice as Times Roman, but they’re readable; spacing using paragraph formatting rather than carriage returns is good practice anyway.
But then there’s figures. On a paginated document, large figures look great floating with text wrapped around them, but terrible inline, because page breaks often cause huge blank gaps in the page before the inline image. Same goes for tables. But our accessibility guidance insists that only inline images/tables are acceptable for screen readers – and Word’s accessibility checker does as well, though the requirement is missing in Microsoft’s accessibility documentation.
Is there some way to make floating images (and floating tables) accessible, so that popular screen readers will know where to parse them in the context of the document? I'm expecting that the documents will be shared as an exported PDF, but even for screen readers to work with the DOCX file would be sufficient.
It sounds as though this is a 'flaw' (design fault) in Microsoft Word's accessibility checker, which doesn't understand text frames being used for table and image layout. Both #slugolicious and my organization's accessibility trainer confirm that having images and tables floating with no additional text apart from the caption is actually fine for modern screen readers.
My error was assuming that my organization's standards included 'No errors from Microsoft Accessibility Checker'. That is not the case.
In the recent months, I've seen that many friends (many of them coders) started writing, long, elaborate mails with footnotes for link, meaning that they write paragraphs [1] and then [2] put the links at the bottom [3]. Like this.
[1] www.example.com/1
[2] www.example.com/2
[3] www.example.com/3
I think it is smart and everything, but I don't understand the process of putting the numbers while you write: I put the reference both when I write and when I edit the text, I swap paragraphs, thus I make a mess with numbers.
Is it a common practice used by some community or is there some editor/plugin that automatically puts the right number in footnotes? Is this mutated from Markdown?
definitely is not Markdown, that compile in HTML, so you should see an anchor instead of plain text.
I think could be an habit from people that want to send plain text mail (not HTML encoded) without use URL inside paragraph to keep readability.
Maybe influenced by Markdown emerging community, where square brackets are used for footnotes. Instead of wiki-markup or LaTeX that use curly brackets for footnotes.
For a quick example, check the very own StackOverflow flavored markdown for links.
Maybe you should also check W3C Web Annotation Working Group for some news.
Is there a best way to localize the language settings? Say the situation is that you have already a working application in, say, French, and wanted to completely localize it in English.
So is there any way to easily localize the application while minimizing the impact on the application's code, I mean there mustn't be any major changes to the code itself but adding some would suffice. I've heard of using resource files in VB6 but it seems to have an issue with its fonts specially in Japanese characters, it throws out a garbage chars specially on labels. Now, what's the best way to change the charset of a application without applying too much changes in the applications code.
This application has a legacy code to I have to deal with it.
I use resource files, and replaced (almost) every string in the codebase and UI with an ID.
Whenever I display a string, I then call a single function that takes a string like {#1234} and loads string ID 1234 (using LoadString() and returns it.
For the UI, I enumerate every control on the form and pass the visible strings to the same function.
This meant a single call to Localise Me in the Load event of each form and a TranslateString("{#1234}", "name", Name) whenever I set/display something dynamically.
For the fonts, see this example from the Visual Basic Programmer's Guide. I call this on every control as part of the Localise method.
Don't forget that different fonts and languages take differing amounts of vertical space for the same text. The form layout also needs to be adjusted to take this into account, or reflow dynamically (align controls to longest label, shift down to allow for longer full width text, etc.)
Regarding the "while minimizing the impact on the application's code" part of the question, I would suggest encapsulating the resource-lookup-related features into a class of its own, exposing methods to fetch a string (by passing a culture specifier argument).
Then the bulk of the work is to convert your code from perhaps hard-coded strings to method calls to retrieve the appropriate strings. Implementing a variant of this answer: Implementing String.Format() in VB6, could greatly simplify your life if you also encapsulate the notion of 'culture' and introduce some cultureInfo argument to this StringFormat method's signature (perhaps call it StringFormatLocal).
The point being, if the current app isn't localized, it's probably not concerned about localization; localizing it means introducing a new concern so in order to affect legacy code as little as possible, you'll need to seek & destroy "magic strings" and "magic formats" all over the code base (and if your UI has hard-coded design-time captions and images, remove those as well), replacing them with calls to your localization API - keeping the localization concern separated.
I'd like to see another answer here with more details about storing, loading and especially using non-ANSI string resources...
Is it possible to customize Recaptcha to display text in English only words?
As recently, I found that text could be displayed in other language like Hebrew.
Here is example:
To be honest it is not possible to type such words for ordinary users having keyboard with Roman alphabet and not many know that image can be redrawn.
AFAIK, via the API you may only customize the interface, not the images.
reCAPTCHA uses scans from the real books, so sometimes, even in latin books there are some non latin characters too.
But there should be no problem here. reCAPTCHA displays always two words: one unknown even for the reCAPTCHA (probably the Hebrew in this case), and the other one, which is really checked.
So the user may misspell the Hebrew, but it's OK when he types the other one (latin) word as expected.
(Only guesses, but I think that's how this thing works).
Have you looked at: http://code.google.com/apis/recaptcha/docs/customization.html#i18n?
That's the API. It talks about setting the translation, but I've never used it, so I'm not 100% sure whether it can do what you want.
Due to where Recaptcha gets its captcha string from (text scanned from books), it could very well be limited to languages that use the Latin alphabet.
I bet that Reviled is the challenge word (the one scanned from the book) and the other is the test word (the one it uses to verify whether or not the person who typed the challenge word is actually typing something legitimate or not).
I run a site where it is important to have good and simple URLs that need to be localized.
Example for the english version:
example.com/car/?type=fiat
Example for the Swedish version:
example.se/bil/?typ=fiat (bil is car in swedish)
And ofcourse I would like to handle all of these URLs from the same codebase. What is the best way to handle this?
Should I set up several controllers (CarController, BilController) or is there a "cleaner" way to handle localized controller names?
BR
Niklas
Don't do that. Ever.
Microsoft, a really big, powerful and resourceful company tried that with Excel. In English versions of Excel, you use IF() in formulas. In the German version, it's WENN(). In French, it's QUAND(), I think. In Japan, it's probably ば(). Now imagine someone from Japan sends me an Excel sheet ... There are two options:
"I'm sorry, I can't open this file"
Translate all names on the fly
Doing #2 seems simple enough ... until you run into a word which uses the same letters but has a different meaning in two languages. Example "see". Means "look" in English and "lake" in German. Since you don't know all the languages in the world, you have no chance to figure out which collisions you will have before it is too late.
Also, how do you know which name to use? From the language in the browser? Or do you hate your international customers who occasionally use the Swedish main site? How do you handle Asian languages? Will the URL be server/%E6%AC%80%E6/?%AD%81%E6%AB=fiat?
Don't. Do. That. Ever.
What about rewriting the URL depending on the domain? This way, the Zend framework will get only the English names, while the URL can use localized names.