PDF Creation

[20_MuleTeam_Borax]

I need to create some PDF files. I think the proper program is Adobe Capture. Unfortunately, it isn't free. I looked for free linux Adobe Capture-compatible software @ freshmeat, but was unsuccesful. Does any exist?

thx

[demian]

There are loads of programs to make pdfs from ps or dvi files. Will that help? I'm gonna take a look for the links but I think they are included in some LaTeX distros...

[Bradmont]

Why on earth would you want to create PDFs? There are too many of them as it is. PDF is a horrible format! What's wrong with text? Or HTML? Hmmm?
So I'm a minimalist. Sue me ;).

[The King Ant]

If you want to make just a simple text pdf, you can type it in abiword and save it as a ps file, then convert that to pdf.

If you want graphics and stuff... I dunno, you might have to buy the adobe program.

I'm not really fond of pdf, but it can be useful for some things.

[Ryeker]

The Adobe program is Acrobat (unless they changed it on me...).

PDF is really nice because unlike text and HTML, it looks EXACTLY the same on all platforms, be it Linux, Windows, Mac, etc. Text? No graphics or formatting, the difference with UNIX and Windows formats. HTML? Heh, you all know about how different pages can look just on different computers with the same operating system with different browsers.

[klamath]

Why on earth would you want to create PDFs? There are too many of them as it is. PDF is a horrible format!

What's wrong with PDF? Plain text just doesn't cut it sometimes. I usually write my school work in LaTeX, produce a PDF file, and then I know I can print out or view the PDF file from almost any computer, and it will look exactly the same. There are plenty of Free Software ways to create and view PDFs...

[Bradmont]

If you want something that can be viewed identically in any platform, make an image ;). The main problem with PDFs is that they're way too big. A friend of mine converted a 4 meg PDF into a 60K of HTML and jpegs, without loosing anything, and it was viewable on most browsers without errors. (He is quite skilled with HTML and it was a fairly exeteme case for him to be able to decrease the size so much, but if HTML is browser dependant, that is just a sign of incompetence.)
I said it earlier, but I'm minimalist. If it can be done in a more compact or more efficient way, it should be. The single argument that I will agree to for PDFs is printability... but that'w what postscript is for ;).

[Strike]

Bradmont, you lose quality when you do that, I believe. PDFs scale a lot better than most graphic formats (because it uses Postscript, I believe). I personally do all my work in LaTeX and then export to Postscript files.

Also, I think Adobe Distiller can produce PDFs as well, and it may be cheaper than Acrobat - I forget, since I never pay for software when there's a better free alternative.

[StanLin]

Gnumeric has a "save as PDF" option.

[Molecule Man]

There are uses for PDF. At work we have several hundred if not thousands of manuals in electronic format. The old ones wer produced in watermark, and they were esentially jpeg files. The newer ones were in pdf format. The pdf files were much easier to read. Even the old manuals which have since been converted to pdf, but lack the searching features are better.

Anyways the command is ps2pdf original_postscript_file target_pdf_file

Enjoy.

[arioch]

How do I drastically decrease the size of a PDF on GNU/Linux?

[ph34r]

Print to a ps file, then use ps2pdf

Or do it all in Open Office

[arioch]

Well ok, since you answered THAT fast, I reopened the issue. ;)

I already looked in OO, I can't see where to to achieve the desired result and OO also getls all "clunky" and symbolweird-acting when I try to re-open the file with it.

[ph34r]

PDFs are print information. Use your original doc, then export as PDF.

[arioch]

The original file was a PDF file itself (A signed contract) that I printed out, signed myself, scanned back into the machine and turned into a document in OO Draw that I could convert into a PDF. A PDF that is now significantly bigger than the first. I can't see how I can decrease the size of the resulting PDF from the OO Draw file I made of it through.

[bwkaz]

You can't. You started with a fairly-small PDF that used actual characters (usually one-byte, or possibly two-byte, entities) to represent each character on the page. When you scanned it in, you turned each of those characters into pixels -- you now need a huge 32x32-pixel chunk of data (depending on the DPI) to represent the character. The pixels are probably represented as full-color, so at least 3 bytes per pixel. That comes out to (uncompressed) at least 3K worth of data. Compression can help a bit, but you're already 1500 times bigger than when you started -- compression can't get you a 1500:1 ratio.

So obviously your resulting PDF is going to be a heck of a lot bigger. There's not really anything you can do to fix that, short of avoiding the print-and-scan operation, and getting your signature into the PDF some other way. (PDFs should be able to contain an image on just part of the page. You could scan your signature (don't need to sign the actual contract) and dump it into the original page at just the right point. This also points out a huge problem with sending "signed" contracts back to whoever in an electronic format -- who's to say you actually signed the contract? All the recipient knows is that they have something with your signature on it, but there's no guarantee that you actually put it there...)

There may be some kind of optical character recognition program that claims to be able to turn each of the 32x32-pixel chunks back into one or two bytes, but I'm not sure how well they'd work. Especially on Linux; maybe tons of open-source OCR programs exist, but if so, I haven't seen them.

We have the exact same problem at work -- people print out 5 or 6 different documents (each at maybe 100K), carry the paper to a copier, scan them in and email the PDFs back to themselves, and then forward that email out to a customer or supplier. And when the now-10-meg message gets bounced because it's too big, they (of course) just try it again. Then when the second message also gets bounced, they complain to us. Like we can change the customer's (or supplier's) email system to allow a larger message... sigh.

Sorry, I'm bitter about that one... ;)

[DrChuck]

I agree with bwkaz, your problem was that the scanned pdf was just one big image, while the original contained actual characters. You may find that you can't select text in the scanned pdf, while you could in the original.
I googled around ( like most people following this thread), and found a method based on gimp, which I tried, and seems to work to some degree, although it is a bit clunky. http://www.wikihow.com/Edit-PDF-Files-in-Linux. You want to open your pdf with gimp, paste in an image of your signature, then save the result. I'm not sure of the best option for the last step - you can save as an image and print to pdf, or save as postscript and convert to pdf. Whatever I tried, the quality didn't seem as good as the original, but I'm no gimp expert by a long shot. The size of the modified file was only a little bigger than the original, as desired.

[irlandes]

OpenOffice, at least the 1.x versions include EXPORT as PDf.

LyX also allows printing as PDF. Back in Mandrake 8.x days, I sat day after day in my room in our house in Mexico City typing in my letters from the Army to my family. I printed as pdf, and photocopies on acid-free paper, and gave an unbound book of that book to each of my children.

LyX allows certain commands, such as makeindex, which in the end, except for the expensive binding, allows a book to be printed on your inkjet that looks exactly like a commercial book.

My son is a college research prof, and has to publish in math journals. He uses LaTeX, line at a time. I am lazy, so I use LyX.

[JohnT]

You could try the OCR method there are some good ones available for Open source. One of the better ones has been in development for a "mighty long time". read about it here (http://www.linux.com/articles/57222)
and download it here (http://sourceforge.net/project/showfiles.php?group_id=158586)