> W dniu 2011-03-02 22:58, a.l.e pisze: >> > and gordom can select the text, so the glyphs are not converted into >> > vectors. >> > >> > as i wrote, at some time there were some improvements with the way >> > glyphs are placed in the PDF... so, gordom, please update to 1.4RC1 >> > and test it there! > 1.4RC1 didn't help at all:-(. It's even worse. When the document is > exported to .pdf directly from Scribus, the text can't be recognized > properly and therefore a search option doesn't work (in 1.3.7 it was). > The text can't be highlighted also. So there is a regress at this point. I am using 1.4RC and have had no problems with creating PDF's and then shrinking them via the compress _newsletter script http://wiki.scribus.net/canvas/Web_optimised_PDF, AND keeping searchable text in the PDF.
What I do know is that sometimes, the compress script doesn't properly compress the PDF, leading to saving the text as a non searchable text. The text on the PDf is also formatted with a slight 'smudge'. I have found that this occurs when you save the original Scribus PDF in greater than PDF 1.3 format, ie PDf1.4 or higher. If you always save in PDF 1.3 from Scribus the problem doesnt occur. I think the reason why the compress script and gs in particular fails to create the type of indexable PDF you want is that if you have transparent images, gs has a problem with this, and fails to render the text as text from this point. I get a good page, but then 'bad ' pages afterwards if there is a transparent image , ie a png on the page. By saving in PDf 1.3 format, it wont allow transparencies anyway, so it stops gs trying to deal with transparent objects. I'm sorry if this is a little vague but it's more from what happens rather than why or how, but if you keep to saving in PDF 1.3 then I *think* you will be ok. I've also used thsi approach from 1.3.3.13 , thru 1.3.5 and up and it always works for me -- Kind Regards, Julian UK
