Annotation of /psiconv/trunk/docs/html4

This README provides information about the psiconv HTML4 output generator.

Output files generated using the option `-t HTML4' use cascading style sheets
(CSS1) with embedded style rules to specify all text formatting information,
rather than using the more commonly-used HTML version 2.0 and 3.2 elements such
as <FONT>, <B>, <CENTER> etc.  Output files can thus only be properly displayed
on the more recent web-browsers such as Netscape 4.x and IE 4.0, but the output
is a much more accurate conversion from the original Psion document than is
possible with HTML 2.0 or 3.2.  Browsers that do not support cascading style
sheets will show the text but no character or paragraph formatting.

Output files that do not contain user-written HTML constructs should comply
with W3C's HTML 4.0 Strict DTD [not checked yet; anj 20-Jun-1999].

The text on the first line of the document header is used as the page title.

Hard page-breaks are converted into a horizontal rule <HR>.  Unfortunately in
Netscape 4.51 any text following in the same paragraph loses its styling, so
this should really only be used at the end of a paragraph.

Paragraph borders using the types dot-dash and dot-dot-dash are converted to
dashed and dotted respectively, as these two Psion types do not have a direct
equivalent in HTML 4.0.  However Netscape 4.51 only seems to display borders as
solid, and in only one color (black).

Netscape also doesn't seem to handle superscript and subscript properly using
the style sheet approach, so the reason these don't work is due to their bug,
not mine.

Bullets are supported, but the output is not quite the same as on the Psion.
HTML lists don't allow you to pick your own character for the bullet like the
Psion does (you could use an image of the relevent character, but you'd have to
create it on-the-fly to get the right foreground color for full support), and
they're not particularly straight-forward to use, so I don't. 

Short HTML constructs such as hyperlinks and images can be entered in the Word
document by using the Psion characters CTRL+139 and CTRL+155 (the single
angle-quote characters lsaquo and rsaquo), which are converted into < and >
respectively in the output file.  Longer constructs such as tables would
probably interact with the paragraph formatting to the detriment of both,
although if kept "on one line" (ie within a single Psion paragraph) they may be
feasible.

The only major facilities provided by Psion Word that are missing from the
HTML4 output are support for tabs and embedded objects.  Tabs look to be
impossible to implement properly because HTML does not provide an equivalent
construct (tables are not suitable for anything other than superficial support.
The classical typewriter model of tabs needs to know the current print position
in order to work out which tab stop to align to next; the print position of a
particular character will depend upon the font in use).  Embedded objects need
the relevent stream formats to be documented and suitable converters
implemented (embedded Sheet files could be converted into tables, and Sketches
into inline images).

- Andrew Johnson <anjohnson@iee.org>
1	frodo	2	This README provides information about the psiconv HTML4 output generator.
2
3			Output files generated using the option `-t HTML4' use cascading style sheets
4			(CSS1) with embedded style rules to specify all text formatting information,
5			rather than using the more commonly-used HTML version 2.0 and 3.2 elements such
6			as <FONT>, <B>, <CENTER> etc. Output files can thus only be properly displayed
7			on the more recent web-browsers such as Netscape 4.x and IE 4.0, but the output
8			is a much more accurate conversion from the original Psion document than is
9			possible with HTML 2.0 or 3.2. Browsers that do not support cascading style
10			sheets will show the text but no character or paragraph formatting.
11
12			Output files that do not contain user-written HTML constructs should comply
13			with W3C's HTML 4.0 Strict DTD [not checked yet; anj 20-Jun-1999].
14
15			The text on the first line of the document header is used as the page title.
16
17			Hard page-breaks are converted into a horizontal rule <HR>. Unfortunately in
18			Netscape 4.51 any text following in the same paragraph loses its styling, so
19			this should really only be used at the end of a paragraph.
20
21			Paragraph borders using the types dot-dash and dot-dot-dash are converted to
22			dashed and dotted respectively, as these two Psion types do not have a direct
23			equivalent in HTML 4.0. However Netscape 4.51 only seems to display borders as
24			solid, and in only one color (black).
25
26			Netscape also doesn't seem to handle superscript and subscript properly using
27			the style sheet approach, so the reason these don't work is due to their bug,
28			not mine.
29
30			Bullets are supported, but the output is not quite the same as on the Psion.
31			HTML lists don't allow you to pick your own character for the bullet like the
32			Psion does (you could use an image of the relevent character, but you'd have to
33			create it on-the-fly to get the right foreground color for full support), and
34			they're not particularly straight-forward to use, so I don't.
35
36			Short HTML constructs such as hyperlinks and images can be entered in the Word
37			document by using the Psion characters CTRL+139 and CTRL+155 (the single
38			angle-quote characters lsaquo and rsaquo), which are converted into < and >
39			respectively in the output file. Longer constructs such as tables would
40			probably interact with the paragraph formatting to the detriment of both,
41			although if kept "on one line" (ie within a single Psion paragraph) they may be
42			feasible.
43
44			The only major facilities provided by Psion Word that are missing from the
45			HTML4 output are support for tabs and embedded objects. Tabs look to be
46			impossible to implement properly because HTML does not provide an equivalent
47			construct (tables are not suitable for anything other than superficial support.
48			The classical typewriter model of tabs needs to know the current print position
49			in order to work out which tab stop to align to next; the print position of a
50			particular character will depend upon the font in use). Embedded objects need
51			the relevent stream formats to be documented and suitable converters
52			implemented (embedded Sheet files could be converted into tables, and Sketches
53			into inline images).
54
55			- Andrew Johnson <anjohnson@iee.org>
frodo@frodo.looijaard.name	ViewVC Help
Powered by ViewVC 1.1.26