Well I agree with your comments and surely on pdfjs being fragile.<div><br></div><div>BTW we have mentionned earlier the Html5 version of FlexPaper. I guess that the drawbacks of this version are different from those of the flash one. Shall we contact the company to see how we can give it a try? </div>
<div><br></div><div><br></div><div>Loïc <br>
<br><br><div class="gmail_quote">On Thu, Jan 19, 2012 at 5:25 PM, Nicolas Limare <span dir="ltr"><<a href="mailto:nicolas.limare@cmla.ens-cachan.fr">nicolas.limare@cmla.ens-cachan.fr</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi,<br>
<div class="im"><br>
> I would add one moderation to that assessment: unless I'm mistaken, some<br>
> pdf that you tested are not representative of what we expect in<br>
> IPOL.<br>
<br>
</div>Oh, yes. I took a few PDFs I had on my disk and tortured the Flash and<br>
PDFjs readers, to get an idea of their robustness. Only a few of these<br>
articles look like realistic IPOL articles, and none was produced by<br>
myself from the current class. And as long as articles are "somehow"<br>
readable online and available with high-quality downloads, I think<br>
the issues can be worked around.<br>
<br>
The only problem that may be serious is the problems encoutered when<br>
rendering some graphics, which could be happen in our PDFs.<br>
<br>
My impression is that the javascript reader is too experimental and<br>
fragile now to be used on the public pages of a journal. This would<br>
leave us with:<br>
* PDF plugin and Flash fallback<br>
* full-text, as good as possible, but probably not very good<br>
it will require a big WARNING! see also this article abourt the<br>
conversion: <a href="http://kwarc.info/kohlhase/papers/mcs10.pdf" target="_blank">http://kwarc.info/kohlhase/papers/mcs10.pdf</a><br>
* downloadable PDFs<br>
<div class="im"><br>
> <a href="http://dev.ipol.im/~simonl/php/pdfInHtml/pdfjs/viewer.html?file=../upload/Carbonell_-_On_Man-Computer_Interaction_:_A_Model_and_Some_Related_Issues.pdf" target="_blank">http://dev.ipol.im/~simonl/php/pdfInHtml/pdfjs/viewer.html?file=../upload/Carbonell_-_On_Man-Computer_Interaction_:_A_Model_and_Some_Related_Issues.pdf</a><br>
> -> PDFjs: uses lots of computing resources<br>
> -> Flash: doesn't render anything<br>
> """<br>
> This file is not really relevant, since I hope we are not going to<br>
> accept scanned pdfs :)<br>
<br>
</div>Sure. But we may have problem when an article has a lot of large<br>
images, which may pose similar problems.<br>
<div class="im"><br>
> """<br>
> <a href="http://dev.ipol.im/~simonl/php/pdfInHtml/FlexPaper/index.php?doc=../upload/docs/stronglinks-ssdbm2010.swf" target="_blank">http://dev.ipol.im/~simonl/php/pdfInHtml/FlexPaper/index.php?doc=../upload/docs/stronglinks-ssdbm2010.swf</a><br>
> -> Flash & PDFjs: very bad math rendering page 4 (bottom)<br>
> -> Flash: abnormal gray marks on page 3 when the illustration at the<br>
> top of the page is out of the window<br>
> -> PDFjs: abnormal color in the illustration at the top of the page 3<br>
> """<br>
> I guess the maths rendering problem is related to an uncommon way of<br>
> generating the math (not latex?).<br>
<br>
</div>$ pdfinfo stronglinks-ssdbm2010.pdf<br>
Creator: TeX<br>
Producer: pdfTeX-1.40.9<br>
CreationDate: Mon May 10 23:18:12 2010<br>
ModDate: Mon May 10 23:18:12 2010<br>
Tagged: no<br>
Pages: 18<br>
Encrypted: no<br>
Page size: 612 x 792 pts (letter)<br>
File size: 1721022 bytes<br>
Optimized: no<br>
PDF version: 1.4<br>
<br>
-> it's pdfTeX, with embedded fonts.<br>
<div class="im"><br>
> """<br>
> <a href="http://dev.ipol.im/~simonl/php/pdfInHtml/pdfjs/viewer.html?file=../upload/IPOL.pdf" target="_blank">http://dev.ipol.im/~simonl/php/pdfInHtml/pdfjs/viewer.html?file=../upload/IPOL.pdf</a><br>
> -> PDFjs: the logo is not resterized with a low resolution, it is<br>
> vectorial in the PDF file<br>
> <a href="http://dev.ipol.im/~simonl/php/pdfInHtml/FlexPaper/index.php?doc=../upload/docs/IPOL.swf" target="_blank">http://dev.ipol.im/~simonl/php/pdfInHtml/FlexPaper/index.php?doc=../upload/docs/IPOL.swf</a><br>
> -> Flash: dirty jpeg rendering of the logo (JPEG Gibbs artifacts at<br>
> 500% zoom)<br>
> """<br>
> This would be a bigger issue. Do we really need a vectorial logo in the<br>
> articles? At least in the web rendering version, I would say that a bitmap<br>
> version should be sufficient. And we could still use the vectorial one in<br>
> the distributed pdf.<br>
<br>
</div>Yes. But this probably means that some other vector graphics will be<br>
rasterized by the Flash conversion, and we loose the possibility of<br>
using the zoom to explore details. Not a terrible problem, but not as<br>
good as native PDF rendering.<br>
<div class="HOEnZb"><div class="h5"><br>
--<br>
Nicolas LIMARE - CMLA - ENS Cachan <a href="http://www.cmla.ens-cachan.fr/~limare/" target="_blank">http://www.cmla.ens-cachan.fr/~limare/</a><br>
IPOL - image processing on line <a href="http://www.ipol.im/" target="_blank">http://www.ipol.im/</a><br>
</div></div><br>-----BEGIN PGP SIGNATURE-----<br>
Version: GnuPG v1.4.10 (GNU/Linux)<br>
<br>
iEYEARECAAYFAk8YQ/IACgkQvviFAPpCP08rxACfWXTPgQlfpuAL6klbnbeLKIzd<br>
rbEAoJmxtBXp4szKb2tNsIqd8exX9eyL<br>
=u3P6<br>
-----END PGP SIGNATURE-----<br>
<br>_______________________________________________<br>
discuss mailing list<br>
<a href="mailto:discuss@list.ipol.im">discuss@list.ipol.im</a><br>
<a href="http://tools.ipol.im/mailman/listinfo/discuss" target="_blank">http://tools.ipol.im/mailman/listinfo/discuss</a><br></blockquote></div><br></div>