Content-Type: text/html At 20:57 98/06/10 +0200, Jacob Palme wrote: > At 08.35 +0200 98-06-10, Martin J. Duerst wrote: > > What I wanted to express in my mail (and probably didn't do > > well enough) is that these specifications are not enough, that > > most probably some implementor advice is needed, because there > > seem to be different practices for the size of the "normal font" > > for mailers and browsers, and this leads to interoperability > > problems between users that use separate tools for mailing and > > browsing and users that use the same tool for these activities. > > Is this a particular problem for MHTML? Is this not an issue for > those who make the standards for HTML? My claim is that it is very much an issue for the interaction between email and web browsing, which is at the core of the MHTML work. > Take the following HTML sequence with relative sizes: > small text
>
normal text
> normal text
> big text
> > And compare them to the following HTML sequences with absolute sizes: > small text
> normal text
> normal text
> big text
> > Both Netscape and Explorer will render the upper sequence > with relative sizes in the same way as the lower seuqnce > with absolute sizes. Thus, there is a built-in assumption > in both Netscape and Explorer that the relative sizes are > offsets from the absolute size "3". Is this not specified > in the HTML standard? This is not the problem. The problem is the incompatibility of the following two scenarios: SCENARIO A ========== I use Eudora for mailing and Communicator for browsing. I use both of them with their built-in default () font sizes. This is 9pt for Eudora, and 12pt for Navigator. 9pt is very nice for reading mostly plaintext email. 12pt is very nice for looking at web pages with colors, links, and so on. These defaults make sense. SCENARIO B ========== Somebody else uses an integrated tool (i.e. Communicator for both email and browsing). This tool only has a single default font size. The user also wants to look at e.g. mail with 9pt, and web pages with 12pt. To achieve this, s/he somehow sets mail to or . When this arrives at my end, it's interpreted as "somewhat smaller than 9pt", which is too small to read, instead of "somewhat smaller than 12pt, which is what the sender intended. > If you get an e-mail with very small, almost unintelligble > characters, this could have two reasons: > > (a) The sender had set his HTML editor to use a very large > base font, for example to set HTML size "3" equal to 18 pt > font size. The sender had then written the text with " SIZE="-2"> to get it looking good for him. If this is the > case, the sender had set the preferences in his HTML editor > wrongly. > > (b) The recipient had set his web browser to use a very > small base font, for example to set HTML size "3" euqal to > 9 pt font size. If this is the case, the recipient had set > the preferences in his web browser wrongly. The above is written on the assumption that there is only one preference for all of email and browsing. My claim is that because these are different modalities with different histories, this assumption is in many or most cases not true. I do not want to set the default font size for my email higher than 9pt, because that's the best size for plain text email. The sender does not want to set the default font size for browsing smaller than e.g. 12pt, because this again is an ideal size for browsing. Both of us want to see email (even if it's sent in HTML format) in a size appropriate for email, namely something around 9pt. Currently, implementations produce something else. An advice to implementations may be necessary to solve this. Regards, Martin.