Re: [WSG] Strange character encoding issue
On Wed, Nov 19, 2008 at 11:05 AM, James Jeffery [EMAIL PROTECTED] wrote: Never had a problem with character encodings on web pages, but since I reinstalled the OS on my iMac I have had an issue. Some of my characters, especially when using ' seem to mess up. This is the page, content and layout are simple as it's for a uni assignment: http://mi-linux.wlv.ac.uk/~0802390/overview.html Check out the overview.html page, and notice the issues. There is one noticeable in the overview page ‘SOAP’ Your HTTP header declares the encoding to be ISO-8859-1 while the HTML (and presumably your text editor) think it is UTF-8. The HTTP header trumps all other sources of encoding information. If you can't change or silence this header on your server, you'll need to save your pages as ISO-8859-1. You might find this article about encoding sources interesting: http://NikitaTheSpider.com/articles/EncodingDivination.html -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Strange character encoding issue
On Wed, Nov 19, 2008 at 11:32 AM, James Jeffery [EMAIL PROTECTED] wrote: I don't own the server. You don't need to own the server to be able to alter its behavior. Many (most?) ISPs allow you to customize aspects of your site. Anyway. I saved as ISO-8859-1, and it works on windows now but not on Mac. Pulling my hair out at this issue. We might be able to help if you're more specific. Windows is an OS; what *browser* are you using? What does that browser tell you it thinks the page's encoding is? Some of the characters you used (like the curly quotes) might not be in the ISO-8859-1 character set. Browser makers are aware of this, and even if the encoding is declared as ISO-8859-1, the browser will treat the encoding as Win-1252 which is an ISO-8859-1 superset. Wikipedia can tell you the difference. On Wed, Nov 19, 2008 at 4:23 PM, David Dorward [EMAIL PROTECTED] wrote: James Jeffery wrote: Never had a problem with character encodings on web pages, but since I reinstalled the OS on my iMac I have had an issue. Your server says: Content-Type: text/html; charset=ISO-8859-1 But the data is UTF-8. -- David Dorward http://dorward.me.uk/ *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] *** *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] *** -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Question about presenting numeric percentages and accessibility.
On Wed, Oct 15, 2008 at 10:01 PM, John Unsworth [EMAIL PROTECTED] wrote: Hi all, Just a quick question. I'm writing up a website for a simple brochure site, and the copy I'm provided with refers to something 1/3 of total or colour 2/3 of natural and so on. And it just occured to me, would Number Slash Number (ie; 1/2) cause any issue in regards accessibility, be it screen readers or poor reading or math skills (the correct term for this alludes me for the moment, I'm thinking dyslexia, but not sure that correctly accounts for all potential users). As such I wondered if the abbr tag might be appropriate, or if anyone has a better, more suitable sugestion? Why not enter the Unicode fraction characters directly? It's not easy to enter them from they keyboard, so what I usually do is enter them as references (e.g. frac34; or #8532; as Todd B. suggested) and then copy paste them from the browser window into my text editor. It makes the source HTML much easier to read. Here's an example. The section called Encodings has a 2/3 in it, and there's some other fractions floating around in there too: http://NikitaTheSpider.com/articles/ByTheNumbers/fall2008.html HTH -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Looking to source a JAWS version
On Wed, Sep 24, 2008 at 4:13 AM, David Dorward [EMAIL PROTECTED] wrote: Thierry Koblentz wrote: http://www.freedomscientific.com/fs_downloads/jaws.asp I don't know about the demo version on that page, but they used to offer a full version that would work for 30 minutes at a time (you needed to reboot the computer after 30 minutes if you wanted to use it again). With, I believe, a license that explicitly forbids using it for the purposes of testing websites for screen reader compatibility. It's true, or at least it was when I looked at the license about a year ago. We switched to using Window-Eyes for testing for just this reason. -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Encoding odities
On Sun, Jul 13, 2008 at 2:49 PM, Mordechai Peller [EMAIL PROTECTED] wrote: David Hucklesby wrote: FWIW - The META content-type is only relevant to pages read from a local file-- for example, when someone saves your page to disk. Not true. I recently had some non-local UTF-8 files where some special characters weren't displaying properly in IE6. When I added the missing meta tag, the problem was solved. Mordechai, you're correct. The encoding declared in the META tag *can* be relevant, although it is trumped by what (if anything) is specified in the HTTP header. When loading pages directly from disk, obviously there's no HTTP headers involved. You might find this article interesting: http://NikitaTheSpider.com/articles/EncodingDivination.html -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Encoding odities
On Thu, Jul 10, 2008 at 8:27 AM, Barney Carroll [EMAIL PROTECTED] wrote: Hello all, I've got a problem with character set encoding I'd like to rectify. I use UTF-8 as a matter of convenience and ideology, and don't believe it should be that much of a problem. My editor (Notepad++) is set to create new files in UTF-8 without a byte order mark, but when I retrieve files from my server it tells me that they're ANSI. Does ANSI means US-ASCII? The most popular single-byte encodings (ISO-8859-X, Win-1252) and UTF-8 are supersets of US-ASCII, so a US-ASCII file is also valid UTF-8 (and ISO-8859-X and Win-1252) all at the same time. It's pretty easy to write English-language pages that are 100% pure US-ASCII, so this might be your situation. Notepad++ has saved the file as UTF-8, but in this situation that doesn't look any different from US-ASCII (i.e. ANSI). Here's an ASCII chart. There are a lot of things floating around out there that claim to be extended ASCII, Microsoft ASCII, Updated ASCII, etc. None of them are official. ASCII ends at character 127, end of story. http://www.jimprice.com/jim-asc.shtml Here's a list of valid charset names: http://www.iana.org/assignments/character-sets I ran an automatic W3C validation of my markup just a second ago after making some edits and it warns me that no character set encoding was specified (even though the first tag in my heads is meta name=content-type content=text/html; charset=UTF-8). For us to figure out why that is, you'll need to share a URL with us. -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] XHTML 1.1 CSS3 - Is it worth using right now?
On Mon, May 12, 2008 at 10:57 PM, XStandard Vlad Alexander [EMAIL PROTECTED] wrote: HTH wrote: ...server has to do content negotiation in order to send text/html with one doctype (HTML or XHTML 1.0) to IE users and application/xhtml+xml/XHTML 1.1 to everyone else. That means you're generating two copies of all of your content Assuming your are not writing static pages, you only need to generate one copy of content in XHTML 1.1 format and then serve it as any version of HTML as you like. I'm not sure what you mean -- I understand the XHTML 1.1 part, but what do you mean then by serve it as any version of HTML? Are you talking about putting an HTML doctype on XHTML 1.1-formatted code, or serving XHTML 1.1 with the text/html media type, or something else? HTH wrote: Furthermore, content negotiation itself is some work to get done correctly At most, maybe 10 lines of code. Please see: http://xhtml.com/en/content-negotiation/ My point exactly -- that code is not correct. It produces the wrong result when presented with an Accept header of */* which is valid (see http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.1) and indicates that the client can accept application/xhtml+xml. The code is also wrong in that the Accept header can contain preference indicators (q=...). It's valid for a client to indicate that it accept both text/html and application/xhtml+xml but prefers the former. A straightforward substring search won't get the job done correctly. It's true that these are unusual cases and the consequences of getting it wrong are minor (text/html sent instead of application/xhtml+xml). But my point was that it is easy to make mistakes, even if you're getting it right most of the time. There was a recent discussion (pretty vocal, if I remember correctly) on the W3 Validator list about the subject of content negotiation involving people with a deeper understanding and appreciation of the standards than me. You might find it interesting reading. Cheers -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] XHTML 1.1 CSS3 - Is it worth using right now?
On Tue, May 13, 2008 at 3:17 PM, XStandard Vlad Alexander [EMAIL PROTECTED] wrote: Hi Nikita, Are you talking about putting an HTML doctype on XHTML 1.1-formatted code Yes, but normally you would put XHTML 1.1 markup into an template written for a different DOCTYPE as shown in this screen shot: http://xstandard.com/94E7EECB-E7CF-4122-A6AF-8F817AA53C78/html-layout-xhtml-content.gif Hi Vlad, OK, I see what you're trying to do, but the example you provided isn't valid XHTML. If it were, the META tag would have to end in a / and then it wouldn't be valid HTML anymore. In other words, it's a good example of why you can't just change the doctype in order to switch between HTML and XHTML. (In addition, the tags would have to be lowercase if it were XHTML, but that's easy to remedy and also works in HTML.) The (X)HTML in the example and content negotiation code you've suggested is probably adequate (from a practical standpoint) for many Webmasters, but it isn't standards compliant. Given the name of this list, that seems pretty significant. Cheers Original Message From: Nikita The Spider The Spider Date: 2008-05-13 8:43 AM On Mon, May 12, 2008 at 10:57 PM, XStandard Vlad Alexander [EMAIL PROTECTED] wrote: HTH wrote: ...server has to do content negotiation in order to send text/html with one doctype (HTML or XHTML 1.0) to IE users and application/xhtml+xml/XHTML 1.1 to everyone else. That means you're generating two copies of all of your content Assuming your are not writing static pages, you only need to generate one copy of content in XHTML 1.1 format and then serve it as any version of HTML as you like. I'm not sure what you mean -- I understand the XHTML 1.1 part, but what do you mean then by serve it as any version of HTML? Are you talking about putting an HTML doctype on XHTML 1.1-formatted code, or serving XHTML 1.1 with the text/html media type, or something else? HTH wrote: Furthermore, content negotiation itself is some work to get done correctly At most, maybe 10 lines of code. Please see: http://xhtml.com/en/content-negotiation/ My point exactly -- that code is not correct. It produces the wrong result when presented with an Accept header of */* which is valid (see http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.1) and indicates that the client can accept application/xhtml+xml. The code is also wrong in that the Accept header can contain preference indicators (q=...). It's valid for a client to indicate that it accept both text/html and application/xhtml+xml but prefers the former. A straightforward substring search won't get the job done correctly. It's true that these are unusual cases and the consequences of getting it wrong are minor (text/html sent instead of application/xhtml+xml). But my point was that it is easy to make mistakes, even if you're getting it right most of the time. There was a recent discussion (pretty vocal, if I remember correctly) on the W3 Validator list about the subject of content negotiation involving people with a deeper understanding and appreciation of the standards than me. You might find it interesting reading. Cheers -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] XHTML 1.1 CSS3 - Is it worth using right now?
On Tue, May 13, 2008 at 10:02 PM, XStandard Vlad Alexander [EMAIL PROTECTED] wrote: Nikita wrote: the META tag would have to end in a / and then it wouldn't be valid HTML anymore. Sure it would. It may not be in the spec but it's a de facto standard. Even the W3C validator will accept it as valid HTML. I encourage you to try that with the W3C validator. You will not get the result you expect. Original Message From: Nikita The Spider The Spider Date: 2008-05-13 7:49 PM On Tue, May 13, 2008 at 3:17 PM, XStandard Vlad Alexander [EMAIL PROTECTED] wrote: Hi Nikita, Are you talking about putting an HTML doctype on XHTML 1.1-formatted code Yes, but normally you would put XHTML 1.1 markup into an template written for a different DOCTYPE as shown in this screen shot: http://xstandard.com/94E7EECB-E7CF-4122-A6AF-8F817AA53C78/html-layout-xhtml-content.gif Hi Vlad, OK, I see what you're trying to do, but the example you provided isn't valid XHTML. If it were, the META tag would have to end in a / and then it wouldn't be valid HTML anymore. In other words, it's a good example of why you can't just change the doctype in order to switch between HTML and XHTML. (In addition, the tags would have to be lowercase if it were XHTML, but that's easy to remedy and also works in HTML.) The (X)HTML in the example and content negotiation code you've suggested is probably adequate (from a practical standpoint) for many Webmasters, but it isn't standards compliant. Given the name of this list, that seems pretty significant. Cheers Original Message From: Nikita The Spider The Spider Date: 2008-05-13 8:43 AM On Mon, May 12, 2008 at 10:57 PM, XStandard Vlad Alexander [EMAIL PROTECTED] wrote: HTH wrote: ...server has to do content negotiation in order to send text/html with one doctype (HTML or XHTML 1.0) to IE users and application/xhtml+xml/XHTML 1.1 to everyone else. That means you're generating two copies of all of your content Assuming your are not writing static pages, you only need to generate one copy of content in XHTML 1.1 format and then serve it as any version of HTML as you like. I'm not sure what you mean -- I understand the XHTML 1.1 part, but what do you mean then by serve it as any version of HTML? Are you talking about putting an HTML doctype on XHTML 1.1-formatted code, or serving XHTML 1.1 with the text/html media type, or something else? HTH wrote: Furthermore, content negotiation itself is some work to get done correctly At most, maybe 10 lines of code. Please see: http://xhtml.com/en/content-negotiation/ My point exactly -- that code is not correct. It produces the wrong result when presented with an Accept header of */* which is valid (see http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.1) and indicates that the client can accept application/xhtml+xml. The code is also wrong in that the Accept header can contain preference indicators (q=...). It's valid for a client to indicate that it accept both text/html and application/xhtml+xml but prefers the former. A straightforward substring search won't get the job done correctly. It's true that these are unusual cases and the consequences of getting it wrong are minor (text/html sent instead of application/xhtml+xml). But my point was that it is easy to make mistakes, even if you're getting it right most of the time. There was a recent discussion (pretty vocal, if I remember correctly) on the W3 Validator list about the subject of content negotiation involving people with a deeper understanding and appreciation of the standards than me. You might find it interesting reading. Cheers *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] *** -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] XHTML 1.1 CSS3 - Is it worth using right now?
On Mon, May 12, 2008 at 4:42 PM, Simon [EMAIL PROTECTED] wrote: Hi, Does anyone use XHTML 1.1 Of the doctypes that my validator Nikita saw in one sample period, just slightly over 2% were XHTML 1.1. It's worth noting that most, if not all, were sent with the wrong media type. http://NikitaTheSpider.com/articles/ByTheNumbers/#doctypes and does it provide any benefits? Well, compared to what? HTML 4.01 Strict, XHTML 1.0 Transitional or XHTML 1.0 Strict? Is there a reason why not many sites adopt this Doctype and is there any point using right now if your site is 1.0 Strict? One big impediment to using XHTML 1.1 is that it must be sent with the application/xhtml+xml media type which makes IE6 choke. That implies that the server has to do content negotiation in order to send text/html with one doctype (HTML or XHTML 1.0) to IE users and application/xhtml+xml/XHTML 1.1 to everyone else. That means you're generating two copies of all of your content unless you're willing to refuse IE users. Does this sound appealing yet? Furthermore, content negotiation itself is some work to get done correctly, even ignoring the cost of generating both two versions of one's content. Given the extra work required to support XHTML 1.1, there would have to be some pretty darn compelling reasons to use it, and those reasons just aren't there for most people. There's quite enough people who question the use of XHTML 1.0 over HTML (I'm one of them), let alone XHTML 1.1. About XHTML and media types: http://www.w3.org/TR/xhtml-media-types/#summary HTH -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] transitional vs. strict
On Tue, Apr 29, 2008 at 11:06 PM, Hassan Schroeder [EMAIL PROTECTED] wrote: One argument against the use of transitional doctypes is that they're now more than eight years old which makes them about half as old as the Web itself. Do you want to base your site on what was status quo half a Web lifetime ago? Uh, aren't the transitional doctypes pretty much, er, well, exactly, as old as their corresponding strict doctypes? :-) True enough! I said that was a potential argument; I didn't say it was a *good* argument. =) In all seriousness, it sounds like the OP's boss is unconvinced by rational arguments, so why not try some irrational ones? -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] transitional vs. strict
On Tue, Apr 29, 2008 at 2:48 PM, Andrew Maben [EMAIL PROTECTED] wrote: I'm finding myself having to justify my work methods to a boss who has almost zero interest in usability, accessibility or standards. (Though I have managed to get into the long-term plan: ...website that is compliant with W3C standards and Section 508...) One question that has been raised is if site X has pages that validate as transitional, why do you have to produce pages that validate as strict? One argument against the use of transitional doctypes is that they're now more than eight years old which makes them about half as old as the Web itself. Do you want to base your site on what was status quo half a Web lifetime ago? Good luck -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] PNG file sizes
On Wed, Apr 16, 2008 at 1:56 AM, Ca Phun Ung [EMAIL PROTECTED] wrote: Mike Brown wrote: Rachel May wrote: I created the PNGs in Photoshop (CS3) and just wondering if there are any better tools or ways of saving the PNGs for smaller file size, while still retaining their high quality?? http://www.ignite-it.co.uk/ Best. Graphics. Optimiser. PlugIn. I use PNGGauntlet as an after process to optimize those PNGs. http://brh.numbera.com/software/pnggauntlet/ Unfortunately it only supports Windows. The Gimp (graphics editor) has PngCrush built into its save routine. You could also run PngCrush separately, I guess, but I never have because I always use The Gimp. Also, PngQuant runs on *nix boxes (and Windows?) and allows you to reduce the palette size of PNGs. Save a PNG as indexed and then tell PngQuant to rerender the PNG with a fixed palette size (up to 256 entries, I think) and it will dither any colors that don't fit in the palette. You can try with different palette sizes to see what tradeoff of size/quality works for you. Enjoy -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Character Encoding Mismatch
On Sun, Apr 6, 2008 at 12:11 AM, David Hucklesby [EMAIL PROTECTED] wrote: On Fri, Apr 4, 2008 at 4:16 PM, Kristine Cummins [EMAIL PROTECTED] wrote: Can someone tell me how to fix this W3C warning – I'm new to understanding this part. http://validator.w3.org/check?uri=http%3A%2F%2Fwww.beverlywilson.com%2F On Fri, 4 Apr 2008 20:15:19 -0400, Nikita The Spider replied: Kristine, If your server is already specifying the character set (a.k.a. encoding) then you don't need to do so in your HTML. In fact, I'd recommend against doing so, ... The META tag is needed when serving the page from the hard drive - for example, when the page is saved for viewing later. (The hard drive does not send HTTP headers.) That's a good point that I should have mentioned, and I'm glad you brought it up. However, IMO this need is often overstated. Browsers are pretty good at guessing the encoding when they need to. I wouldn't rely on browsers guessing correctly for public pages, but I think the clutter of having duplicate encoding declarations usually outweighs the benefit. Of course, ideally one looks at one's pages using a local Web server. I think Windows Linux come with one preinstalled and I know that OS X does, so this should be within the reach of most folks. Cheers -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Character Encoding Mismatch
On Fri, Apr 4, 2008 at 4:16 PM, Kristine Cummins [EMAIL PROTECTED] wrote: Can someone tell me how to fix this W3C warning – I'm new to understanding this part. http://validator.w3.org/check?uri=http%3A%2F%2Fwww.beverlywilson.com%2F Kristine, If your server is already specifying the character set (a.k.a. encoding) then you don't need to do so in your HTML. In fact, I'd recommend against doing so, and the problem you've experienced is exactly why. If you specify the encoding in two (or more) places, they can get out of synch. You might *think* you're specifying ISO-8859-1 because that's what your HTML META tag says, but if the server says something else, that's what takes priority. It's important to understand that the encoding tells browsers (and other user agents, like Googlebot) how to interpret non-ASCII characters in your page. It's a common mistake to think that these are restricted to accented characters that we generally don't use in English, but content pasted in from Microsoft Word (for instance) is likely to contain non-ASCII as well. In other words, you might be using them without realizing it. If you are, and you get the encoding wrong, then what you see as quote marks (for instance) might look like this to others: †Whatever tool you're using to save files should give you a choice of which encoding/character set to use. You can use ISO-8859-1 to write in English and most Western European languages. Since your Web server is already identifying your pages as such, it might be a good choice. Others have suggested UTF-8 which can represent anything under the sun. That's great, but you'll have to find some way to cajole your server into telling the world that your pages are UTF-8, not ISO-8859-1. If you can't, you'll have to stick to the latter. -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] META content-lang. declared but showing up different (for one person reporting)
On Thu, Mar 20, 2008 at 3:17 PM, Kristine Cummins [EMAIL PROTECTED] wrote: I launched a new site a few days ago and received a report that the site is showing in another language and/or foreign characters even though meta http-equiv=content-language content=en / is declared in the HEAD. That might be useful to search engines for deciding what language your content is in, but I don't think any browsers pay attention to language hints. The person reporting is using Safari (v. unknown) and I think on a Mac (not sure if this info is applicable). is an XHTML site and validated as such. On my Mac using Safari 3.0.4 the site displays just fine. Thoughts appreciated - Thank you, Without any more information, I'd guess the problem is specific to that person's computer. It could be that the font specification of palatino linotype is causing Safari to make some weird substitution that's not working out, but that'd be a bug. Well, there's a bug somewhere... Good luck -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Colour Blindness Statistics
On Nov 11, 2007 8:33 AM, Rahul Gonsalves [EMAIL PROTECTED] wrote: On 10-Nov-07, at 6:33 PM, Gunlaug Sørtun wrote: Rahul Gonsalves wrote: I'm searching for first-hand, authoritative statistics on colour blindness, for use in a formal, academic document. Would anyone be able to point me in the right direction? Hi Rahul, I've also seen the 8% figure frequently cited but, like you, have never been able to find a reference for it. I'd like to know the source(s) for this figure, because I'd like to know what group was assessed. Is that U.S. males? North American males? North American caucasian males? (If the study was done in the USA pre-1950, it wouldn't be unusual to find that it ignored certain ethnic groups.) I would be surprised to learn that the 8% figure is applicable worldwide in all populations. Furthermore, how did the author(s) of the study(s) define color vision deficiency? As I understand it, color vision ability is on a spectrum ranging from tetrachromacy on the high end to complete color blindness on the low end. Where's the cutoff point on that line that defines deficiency? Maybe these are not the right questions to ask. I'm not a color vision expert, or a color expert or even a vision expert. My point is that the 8% figure is too neat to seem real, and I'll bet the reality of the situation is more nuanced and interesting. For anyone interested in the subject, Oliver Sacks' _Island of the Colorblind_ is an enjoyable read. -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Encoded mailto links
On 10/19/07, Chris Knowles [EMAIL PROTECTED] wrote: I noticed this page also uses entity encoding. This is a solution I have used myself but the more I think about it the more I realise realise how ineffective it is really. take the following PHP code: // some page fetching function $html = fetchPage($url); // convert any entites in the page to plain text $html = html_entity_decode($html); now $html contains plain email addresses - with one line of code surely any harvester performs this operation first? Hi Chris, I often see the same argument about Javascript. That is, it is trivial to embed a Javscript interpreter in an email address harvester, so Javascript-protected email addresses are (or soon will be) vulnerable. IMHO, arguments based on the cost of programming the harvester are misguided. Far more important is the CPU and memory cost of running the harvester over the long term. For the harvester, both increased throughput and increased intelligence imply increased addresses harvested. But the former buys them much more simply in sheer numbers -- most addresses are presented unprotected. And the latter (clever harvesting, like running html_entity_decode) is only likely to harvest the addresses of Net-savvy individuals who are the least likely (we hope!) to respond to spam and phishing. One could even argue that a clever harvester is counterproductive in that it will pollute its database with the addresses of these Net-savvy individuals. In short, I think harvesters download HTML pages and run a regex that looks for '@' with text on either side. They care about getting as many email addresses as possible as quickly as possible. Just my $.02, -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Encoded mailto links
On 10/18/07, Anders Nawroth [EMAIL PROTECTED] wrote: Hi! Nikita The Spider The Spider skrev: You might be interested in an experiment I ran that compared a few techniques for protecting one's email address from harvesting bots. The short answer: entity references worked very well I think the time span of your study is to short. I have used the method you used for äcklig, with mixed decimal and hexadecimal numeric entities. In about a year there was no spam, but somewhere at 1.5 years it started a little, and after 2 years there where 100+ spam/day. Hej Anders, That's very interesting, thanks for letting me know! So I think you just push the problem forward, which could be fine in some cases. But when a entity-decoding spam harvester finds the email-address, this will get listed in the same databases as all other emailaddresses. The more traffic your site has, the less difference the encoding will make. I agree. I assumed (wrongly) that the 200+ days of the study was long enough to get found by any harvesters that bothered to decode entities. I'm not surprised to learn, however, that once the address was exposed that it received an ever-increasing amount of spam. This is consistent with my intuition and also what I observed. I think the htaccess-trick linked to by Dejan Kozina looks more promising. I have used this method, but abandoned it because of that some browser wouldn't send the mailto: address to the email client. But this was a few years ago, so this could possibly have changed. This method looks promising to me too but I haven't had a chance to test it yet. -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Encoded mailto links
On 10/18/07, Anders Nawroth [EMAIL PROTECTED] wrote: Ray Leventhal skrev: As a matter of preference, I generally try to eliminate all mailto: links on any site I've been asked to work on. In place, I use a contact form, Me too :-) But then you get form-post spam after a while ... I have begun to add a random token as a hidden field to forms (which is then checked on the server side, ensuring that a form can only be posted once), stopping bots that don't actually read the form every time. (I actually think most of them don't ... when I change the field names in forms the spam form posts continue to use the old ones for a long time.) Mail forms (badly coded ones at least) are also vulnerable to SMTP header injection. Once spammers find a vulnerable form, they can use tools like curl to POST data to your mail processing script without even loading the form. Challenge tokens like yours are a good solution to this problem. FYI, I ran into a nasty problem with Safari related to a challenge string embedded in a logon form. Funnily enough, the solution is to change the field name, as you're doing, Anders. Here's the problem and a solution just to save someone else some headaches. Forgive me, all, if this is getting too far afield for this list. I generated the challenge string server-side and stuffed it into a hidden input field with a name and id of challenge. When I submitted the form I was sent to ProcessLogin.xxx. If I then hit the back button (say because my password was rejected), Safari respected my cache control directives and reloaded the page from the server, thus getting a new challenge string. BUT, once the page was loaded, Safari helpfully repopulated the input fields with the values they held prior to form submission. This was nice in the case of the userid because it saved me some typing, but it caused a problem with the challenge field because Safari *overwrote it with the old challenge string*. The order of events was as follows: 1) I hit back: 2) Safari gets a new copy of the page from the server. This creates a new challenge string that embedded in the challenge field of the page. 3) Safari (re)populates form fields with the values they held when the form was submitted. 4) Safari executes the page's Javascript. What was very confusing was that even View Source showed that the page had the new challenge string but it was the old string that was sent over the wire when I resubmitted the form. (Thank you Wireshark.) Because of the order of 3 4, I could have solved this problem by (re)setting the challenge field in the onLoad event. But that might not fix the problem if there's another browser that has the same quirk but executes those steps in a different order. Instead I came up with the solution of using the challenge string as the field id/name instead of its value. (The value is simply challenge.) Safari's helpfulness leaves this field untouched since it has a new name every time the page is loaded. The challenge string still gets transmitted to ProcessLogin.xxx, I just have to find it by looking through the form field names rather than their values. Note that since the challenge string appears as a form control name and form control names must begin with a letter per the HTML spec, challenge strings have to start with a letter. -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***
Re: [WSG] Encoded mailto links
On 10/17/07, Rick Lecoat [EMAIL PROTECTED] wrote: Hi, can anyone tell me what is the best accessible way (if any) of encoding a mailto: link? I want to make the email addresses on a site usable to screen reader users, but don't want them harvested by spambots. Hi Rick, You might be interested in an experiment I ran that compared a few techniques for protecting one's email address from harvesting bots. The short answer: entity references worked very well and do not need Javascript to work. I am not an accessibility guru so I can't say how friendly that is to screen readers and the like, but I reckon you have plenty of people on this list who can inform you on that score. http://NikitaTheSpider.com/articles/IngenReklamTack.html Hope this helps -- Philip http://NikitaTheSpider.com/ Whole-site HTML validation, link checking and more *** List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] ***