Without questioning the value of this research, I feel the need to ask whether anyone has actually looked at whether the problem is "big enough" to merit the rather drastic changes being proposed.

While I've occasionally seen a few broken chars when few "particular someones" paste from another website, the offense is less serious than the far more common typographical errors that we tolerate from the few fat-fingered members who don't read before submitting. (Several specific persons who frequently post broken chars do so only when copying from a few fairly specific and unique other websites. Even their posts are clean when posting from most sites. I do wonder why.)

If the problem was really significant, I'd expect to have seen complaints in the threads where char errors happen, and I don't find any such objections in any threads I've read.

An additional question is whether changes made now will be valid for the eventual (pending) standards for "web fonts." The W3C committee has "standardized" the "@font-face rule" for using CSS to designate a remote font, and the capability is reported (June 2009) as being available in all major browsers, but most such implementations violate proprietary rights of the font makers whose fonts are used, and the type foundaries are objecting. If a CSS is used, there's still a problem if the machine browsing the site doesn't have the font specified, and at present mudcat likely would be forced to "buy" the specified font in order to put it on the server.

Microsoft has provided for the use of EOT (Embedded Open Type) in which the (encrypted) font is embedded in a document. They haven't had much success with getting people to use it, although IE can, and in some cases does, use it. Other browsers probably can use EOT by now, but it's unclear how much additional burden that use places on website servers.

"All the browsers but Microsoft's, meanwhile, have embraced a technique called "naked" or "raw" font linking, which means
uploading ordinary desktop fonts onto servers." (Technology Review, June 2009) This method clearly, in most cases, violates the foundries' licensing terms, and while they've shown commendable restraint, failure to do "something different" may soon result in lawsuits like the ones over music and videos.

A fairly recent change in copyright rules means that unlike a short time ago a font can be copyrighted as software, and most of the decent ones are ©.

The "Web Open Font Format" (WOFF) is being pursued actively, particularly by a couple of "startup" companies^**, but as proposed now that would require web sites to pay annual fees to lease the use of fonts. That seems to be an unreasonable burden to foist on mudcat when the use, as now, of the fonts on users machines is legal and "already paid for" (if you're using legal copies of the programs that provided them).

^** The startup company called Small Batch is offering WOFF as "Typekit" and another startup named Kernest is offering to "broker" leases. Mozilla has signed on and typographers are circulating petitions for a standard.

Implementation of WOFF also would require significant changes in all browsers, which really means that it won't do a lot of good until our people who are still using Win98/WinME/WinXP buy new computers compatible with the newest browsers.

Clearly defining and describing the particular user practices that result in broken chars now, with the existing mudcat setup, and telling people "how to not do that" would be helpful. The ones who are doing it will ignore the request to not do it, and the rest of us will shrug and go on with our reading and posting.

John