| 1 |
On Tuesday 13 August 2002 18:29, Daniel Mettler wrote: |
| 2 |
> |
| 3 |
> nb. html 4.01 b.1: |
| 4 |
> |
| 5 |
> "Since user agents may vary in how they handle error conditions, |
| 6 |
> authors and users must not rely on specific error recovery |
| 7 |
> behavior." |
| 8 |
> |
| 9 |
> further |
| 10 |
> |
| 11 |
> "For reasons of interoperability, authors must not "extend" HTML |
| 12 |
> through the available SGML mechanisms (e.g., extending the DTD, |
| 13 |
> adding a new set of entity definitions, etc.)." |
| 14 |
> |
| 15 |
> -> it's a wrong bad habit to include browser specific attributes |
| 16 |
> and it's wrong to expect them to be ignored by other browsers. |
| 17 |
|
| 18 |
> this is the wrong view again. i repeat: you simply can not blame |
| 19 |
> browsers for "bad" rendering of not standard compliant pages |
| 20 |
> (i.e. it's *not* a bug). but you can blame them for incorrect |
| 21 |
> rendering of standard compliant pages (i.e. it's a bug). |
| 22 |
> |
| 23 |
> just try to write parsers/renderers without clear specifications |
| 24 |
> and i promise you will miserably fail. i have gone through this. |
| 25 |
> html is rather "adventurous" stuff already - do not let us make |
| 26 |
> it worse. |
| 27 |
> |
| 28 |
|
| 29 |
I stand corrected. I've done some html parsing myself. Did you know that |
| 30 |
altavista relies on not welformed html to do some nifty layout. They put the |
| 31 |
form opening, input, and form close tags each in a different table cell (of |
| 32 |
course this is not welformed) |
| 33 |
|
| 34 |
Paul |
| 35 |
|
| 36 |
-- |
| 37 |
Paul de Vrieze |
| 38 |
Junior Researcher |
| 39 |
Mail: pauldv@××××××.nl |
| 40 |
Homepage: http://www.devrieze.net |
Archives