Paragraph
formats/styles will be preserved best when using the export to
HTML-with-CSS and viewing the output HTML files with a CSS-capable
browser. Style names are filtered so that they contain only alphanumeric
characters, since other characters can cause problems with some browsers.
Since HTML does not support tabs or tab leaders, these
attributes will be lost as Filtrix converts into HTML.
Columns are supported, though not all browsers recognize column
tags.
Tables are converted preserving straddles, rules, captions, and
header and body row identification as completely as is permitted by the
target application.
In rendering paragraph indents into HTML, the CSS filter uses
"text-indent" for first line indents, "margin-left" for either hanging or
block-left indents, and "right-margin" for block right indents. The
standard HTML (non-CSS) filter uses the <UL> construct for block
left indents and "text-indent" as first line indents. The standard HTML
filter cannot preserve the right indent attribute.
Left and right paragraph margins are referenced to a default value of
one inch left and right margins. Margins less than this value are ignored.
Special characters are converted according to the available HTML
character set, which is a subset of the 256 ISO character set. Since this
is generally more limited than most word processor character sets, Filtrix
may have trouble mapping some characters. If Filtrix is unable to
successfully map a given character, it either uses the currently defined
font-face character or, as a last resort, inserts a "?" to indicate that
the character can't be rendered by HTML standards.
Single level autonumbers are supported into HTML, where the
<OL> tag is used. Multi-level autonumbers are written out as literal
text.
Indexing is not presently supported; a later release will
provide an index containing links to each indexed item.
Footnotes: Although there is no explicit HTML footnote feature,
Filtrix creates links for each instance encountered in the source document
and each footnote or endnote is treated as an endnote. A listing of the
footnote contents is concatenated to the end of the document.
Cross-references are treated as hyperlinks.
Paragraph and Character graphic anchor types are created.
Table borders are either on or off, and Filtrix makes this
determination based on whether or not any of the outer table borders are
specified as visible. Fill color is supported in table cells:
image-fills are not supported. Horizontal and vertical straddles,
graphics, and text attributes are supported.
Any hidden text converted from a source document will be written
to HTML using the comment tag.
Notes on
converting HTML to other formats
Autonumbers at one level are created,
interpreting the HTML <OL> tag.
When Filtrix exports HTML to other formats, comment tags and other
HTML-unique tag data will be stored as hidden data if the target format
and filter support this feature.
If the HTML document utilized CSS formatting, the converted
document will use the paragraph oriented CLASS tags as paragraph style
identifiers. For those HTML documents that do not contain CSS, in order to
facilitate editing of the converted document, Filtrix takes each HTML tag
that initiates a new paragraph and turns it into a format tag prefixed by
"HTML_" in the target application. For example, a <P> tag becomes a
format tagged as "HTML_p" in the target application.
If the target application supports it, hyperlinks are converted to
either cross-references or hyperlinks.
Graphics: Only JPEG and GIF images are converted at present.
Only the first image of an animated GIF is converted. Other graphic types
such as PNG will be supported in a later release. Graphic images
referenced by an HTML document must be in the same drive/directory as the
HTML parent file, or may be in either the absolute or the relative path
indicated by the image tag in the HTML document. If image files are not in
one of these locations, they will not be converted. Note: Filtrix will not
convert JPEG to TIFF.
Though HTML does not have headers/footers, when importing HTML
Filtrix does create one all-page header containing a date-stamp and the
file name. It also creates an all-page footer with the page number in it.
For specific details about bidirectional conversions between HTML and
FrameMaker, go to