[XeTeX] [OT?] "Optimized" PDFs?

[Ross Moore]

Hello Michael,

On 17/01/2008, at 4:23 AM, Michael B. Trausch wrote:

> On Wed, 2008-01-16 at 09:55 +0100, Martin Schröder wrote:
>> No, and neither does pdftex. ghostscript comes with pdfopt, which
>> converts normal PDF to linearized PDF.
>
> Thank you for the pointer to the tool; it is perfectly okay by me to
> have separate tools for separate tasks.  :)
>
>> Typically, "Optimized" is used as a synonym for Linearized PDF.
>
> I do find it interesting that optimized PDF files are larger than
> non-optimized PDF files.

They are "Optimized" for starting to display the pages, in order,
in a browser, while the file is still being downloaded.
That is, you can read (some of) the displayed contents of the file
before the whole file is available.

In "ordinary" PDF this is not possible, as the most important
index (which gives byte offsets to important resources, such as
where the data for separate pages can be found) occurs at the
end of the file --- so you need it all before a browser can tell
what needs to be displayed.

Clearly "Linearized" is better terminology, in the sense that
resources are stored sequentially with respect to page order.
However, this may involve duplication of some resources,
such as use of font subsets, rather than a single instance
of the whole font. And there may even be more resource-tree
structures, for each page, than would otherwise be the case.
So yes, files that are "optimized" in this sense are naturally
larger in terms of disk-space required -- but that is cheap
these days.

>
> Is there any truly compelling reason that I would want to optimize my
> PDF documents?

That depends upon how you think your audience will be reading your
files.

> It seems that most documents that you find on the
> Internet---or get from places like EBSCOhost, the ACM, and  
> similar---are
> always optimized.  Does this somehow streamline it for the single case
> where someone is viewing a PDF using, say, Adobe Acrobat Reader as  
> a Web
> browser plug-in?

Yes, that is the main idea.
It might also be easier to read locally on a slow machine,
or one short on disk-space, or when local caching is nearly full.


These are, I think, the ideas involved with Optimization.
Someone else may correct any errors in my explanation.



Cheers,

	Ross


>
> 	--- Mike
>
> -- 
> Michael B. Trausch                                   mike at trausch.us
> home: 404-592-5746, 1                                 www.trausch.us
> cell: 678-522-7934                       im: mike at trausch.us, jabber
> Ubuntu Unofficial Backports Project:    http://backports.trausch.us/
>

------------------------------------------------------------------------
Ross Moore                                         ross at maths.mq.edu.au
Mathematics Department                             office: E7A-419
Macquarie University                               tel: +61 +2 9850 8955
Sydney, Australia  2109                            fax: +61 +2 9850 8114
------------------------------------------------------------------------