Monday, December 12, 2005

Silent Print a PDF !! Print PDF programmatically....

Update: Please do not forget to check the other PDF and print related links in the sidebar


If you have come to this page after a Google search on the following terms - "print PDF in a programmatic method", "silent print", "automate PDF print" then it is probable that you have come to the right place. This would be true especially if you want a cheap / free / open source solution to print PDF automatically.

And if you just want the answer to question on "How to prevent users from saving a PDF document ?", then just go into the Adobe Acrobat forums where the issue has been highlighted as a "most frequently asked question". There is another interesting and well-explained article available on why it is futile to prevent website viewers from saving your "secure" PDF files. And good luck, if you plan to leave already without reading the rest of the article.


First of all, take a deep breath of air; you will be going through quite a long ride now.


Let's go through the basics first.

A printer is a device / resource in the context of operating systems. So all printing is done at the behest of the operating system. For a layman, this would mean - the user has control over what to print and what not to print, how to print and whether the print commands can be archived for printing the same document later on. You might ask - how is it possible to store the print commands ? Well, it is possible to have software / print servers that will store such printing instructions. This is usually the case in my friend's company - the print server software stores all submitted documents for a week, after which they are discarded. Point to be noted : to print something the printer will need printer-level instructions, and those instructions can be stored in printer memory to be reused later.
Coming back to the point, silent printing or for that matter even user guided printing requires code to exercise control over the Operating System. So if you want a silent print without any hassles, your success in achieving it will depend on how much control you can exercise over the Operating System from a given environment - "trusted" desktop application based printing, website based printing, "signed" applet based printing...........................................

Now on to the gory details of PDF.

PDF is a Adobe file format that is a tempered down version of Adobe's Postscript format [technically Postscript is a programming as is any other computer programming language].
So if you want to print a PDF file [with or without user interaction], you will have to convert the information in the PDF file to Postscript. This will enable the printer Postscript driver to issue printer specific commands [note the word "commands", this is one reminder that Postscript is a programming language] that produce the required output accurately.
It is a different issue for printers that come without a Postscript driver. Certain printers especially that of HP might come with PCL [ Printer Command Language ] drivers instead. And there comes the additional step of converting the Postscript data to data understood by PCL. This implies that to print a PDF, you'll have to write code to be capable of converting PDF/Postscript to the native format understood by the printer.
One such product capable of doing that is the free Adobe Acrobat Reader. This well-known product does have it's limitations however. Command line options to print the PDF are undocumented / unsupported by Adobe. The Acrobat Reader process is known to start, print and continue running even after a silent print via command line; a bug that's not yet fixed by Adobe. Which means - you should expect to see a blank Acrobat Reader window after the "silent print". There are "beat around the bush" methods to attain certain amount of success using the commandline options - not for those who want "da perfect solution".
Inter Application Communication (IAC PDF reference file available at link) with the free Acrobat Reader cannot be done via OLE automation give the fact that Acrobat Reader is not a OLE server; you can get some luck only with Acrobat Standard or Pro. Only DDE messages can help a programmer write an application to achieve this. DDE will have restrictions that are not faced by an OLE Server. [If you don't know the difference then you musn't be doing this anyway].

Commercial libraries could alleviate this problem, but they might have to installed at the client systems - licenses for such usage are known to be highly priced [Consider this option in an enterprise application]. Such libraries are available for both .Net and Java. Availability of ports to other languages / frameworks will depend on the demand for the ports.

There are other solutions as well - the Acrobat JavaBean to view and print a PDF file. You can keep away from this, given the fact that it is very obsolete and is known to produce smudgy previews and printouts. Very buggy and should not be used unless your customers don't want printouts. In short, keep away even if this "recent" article" seems to help you.

Rumours have it that Sun Microsystems has an in-house library that produces high quality output. The reason for it's unavailability is however unknown. Seems like there are some licensing issues there, which might also be one of the reasons why the org.jdesktop.jdnc.incubator.rbair packages disappeared all of a sudden. If you got excited by having a look at the Java Print Service (which was initially offered to solve all printing problems!! ) and discovered that there is a PDF Docflavor, then it is time to cool down; take a peek at this link. I'd also quote the most important part of the page (right at the end)........

"Just because the PDF DocFlavor exists doesn't mean you can use it to print a PDF file. When you lookup the PrintService for the PDF flavor, it will report unsupported flavor. That means there is no print service for the flavor. Sun doesn't provide one for PDF files. To the best of my knowledge, nobody else makes one available either. Until such a time as someone does, you can't print PDF files using the new Print Service API."

There are open source solutions as well that might get the job done, but I wouldn't recommend any, given the fact that I consider none of them are mature enough to be used in commercial projects with a heavy stake on print output quality. For some exploratory programming you could try JPedal and PDFBox.

You can otherwise think of buying commercial libraries that will rasterize [convert the PDF data into a bitmap] and print the PDF files. And if you want to avoid all this, you'll need to write your own code to rasterize and print the PDF file - the time penalty for doing such a thing is huge [Commercial rasterization libraries sell for $500 and up for developer licenses, so duplicating that effort will take that many man hours]. Links for exploratory activity include: TallComponents and ICEsoft. The list is not comprehensive though.

I have thus outlined the methods to print a PDF file via a local system application.

Now onto that "silly, but not mediocre" request of silent print of an Internet based PDF file via the browser.
First of all, the browser's job is to display HTML and probably print a HTML page and nothing else. If you want it to print a PDF file, you'll need a browser helper object [BHO]. Try coding BHOs for all browsers and you'll know you're asking for manna.
You could rely on the Adobe Acrobat IE BHO/plugin, but it's not foolproof. If you want to know why, then you ought to read even more carefully now - there are umpteen ways/hacks/methods to disrupt that method. No known document exists, which can pinpoint the Internet Explorer and Acrobat BHO settings that will disrupt the plugin's functionality. Simply turning the plugin off, and switching it back on, by using the Acrobat Reader preferences menu can disrupt that functionality !! So much for relying on Acrobat Reader to be a carefully engineered application.

There is however an easier but hack 'n slash way to do it; use a document level JavaScript print action [ for more information have a look at the Adobe Acrobat JavaScript Guide(pdf) and Reference(pdf) ]. The print action is to execute on any valid action - usually a page action. But remember, JavaScript with Acrobat Reader can be disabled by the user. And it is impossible to close Acrobat Reader [not the opened document] after the print is finished - you'll need to "deliberately crash" Reader after the print is done. And if the PDF file opens in a browser window [ on account of the Acrobat Reader BHO], then you can forget about trying to close the PDF document in a programmatic manner. One of the iText examples shows how a servlet could be written to achieve this. Try a demo here! and make sure you run it in different browsers under different settings (for example : Win XP with SP 2 and without SP2), to understand the varying functionality that it provides. Golden rule to be respected in any software : Never make it behave inconsistently; your users could have trouble describing their problems.

You are lucky if you want an IE only solution with certain restrictions - hop onto ActiveX programming using MeadCo's ScriptX. As again, this solution depends on how stable Internet Explorer will turn out to be. Windows XP SP2 will definitely throw up the security message bar regarding the webpage's attempt to execute code (remember that webpages were to display HTML, not run some virus proof of concept).

Some people also like a "PDF Preview" in the browser. My answer would be - "that's not the job of the browser"; write something to help the browser do that - an ActiveX plugin, Browser Helper Object [BHO] or a Java applet is what you want. The Adobe Acrobat IE/Mozilla BHO might do the job for you if the stars were in the right alignment at the time of your birth; please dont ask me to support the article present at the link. And yes, you can print via the ActiveX plugin / Java applet; but it's not easy given the fact that brilliance always has a limit. ActiveX controls / Java applets that do that task come with the "$$" penalty. You should invest time or money to get that done.

And stop cribbing if you still havent ;-) or ask your boss to stop ;-). Atleast Adobe has licensed the PDF format so that you can use it (naughty boy, you thought everything was free, eh?). Imagine your silly client's fate if you had to write your proprietary file format to enforce a thinly veiled DRM idea of his. Frankly, I've had enough of the people asking for solutions like - " How can I get the client to print my PDF document without saving it. My budget is 50$ ". The answer is - you don't do it with basic code. You'll need to write a DRM plugin (which will cost your client moolah; remember nothing is free) or use Adobe LiveCycle Policy Server or buy security solutions from Adobe's partners, which ever fits into your budget and needs.

And on a parting note, "See ya !! Talk to you later !!!! I hope it was a pleasure to be bored by listening to me, as much as the pleasure I felt in boring you ;-) "


21 comments:

Anonymous said...

There are printers out there which are able to print PDF files directly. So you wont't have to save the PDF on a client, you can get it from a web server directly on paper.

Unknown said...

First, my apologies for my late replies. I hardly noticed these comments lying in the blogger storeroom.

I'll answer the questions one by one...
>>**<<
"Finally, I don't need to print this from an internet file; I have the PDF files. But, I want to call it from my software to print via Unix, and Acrobat doesn't do the command line printing."
>>**<<

My answer would be Ghostcript. It is a tool to print PS/PDF files on Windows and *nix. I did not mention this in my article given the fact that most readers would never have understood GPL, and for that matter even bother understanding how to make Ghostscript work for such a requirement.

>>**<<
"There are printers out there which are able to print PDF files directly. So you wont't have to save the PDF on a client, you can get it from a web server directly on paper."
>>**<<
Yes right. Some printer drivers do support PDF streams natively. I have no idea on how many such printers exist. So I would not recommend trying to send a PDF stream directly to the printer (you will have to write code to send print instructions in an alternate format). It's usually PS/PCL or in the rarest cases of the ancient "ink-stampers", it is even text/lines !!

Anonymous said...

JPS does not make life easy for Java developers. When we implemented printing in JPedal there were alsorts of 'undocumented features', differences between 1.3, 1.4, 1.5,etc.

Java offers both a pageable and printable interface but one needs the printDialog method to be called to initialise it, making it useless for background printing. Its also impossible to extednthe print dialog to add menu options :-(

Unknown said...

Yes, I agree. JPS is nowhere as promising as it sounded at first sight. Eventually, you'll need some software (a service, a plugin or something similar) at the client end to achieve a resonable degree of success in automating the print process.
I've not explored a lot, but I might be right if I state that a printing framework is long overdue; a lot of work needs to be done by the printer driver developers and authors of print format descriptions.

Anonymous said...

Java Printing just creates a large raster and converts that to whatever the printer uses via the driver. Its very inflexible (as far as we can see its still 24bit even if you are printing in monochrome).

We wrote an example of server printing for the JPedal library and put the code into the public domain if you need a worked example.

Unknown said...

I'd like to have a look at it then. Especially if it helps the people visiting this place. Most of them seem to google for server issued printing commands to printers at the client computer. They seem to be doing this in an enterprise environment, or on a LAN, so an applet or smart client based solution would be great, despite the download size and applet startup time.
My friend Pascal Betz implemented a FOP based solution for his requirements, but I would trust JPedal more than FOP 0.91 beta.

Anonymous said...

We want to develop silent printings and another settings related to
printing. We are developing a new active x control for that, and we
would like to know if is possible to send a pdf file to the active x
control.



Is it possible with your libraries? Do you know what could be the solution?

Unknown said...

Hi,
Thanks for the visit. Wish I could call you by name. Anyway, it seems you need to develop a MeadCo ScriptX (http://www.meadroid.com/scriptx/index.asp) based solution. ScriptX is an ActiveX control that is used by several developers to achieve control over the print process on the Internet. It has the ability to print PDF documents of a basic nature from the webpage itself.

And any file can be loaded so long as it can served over the internet by the webserver on which the file is hosted.

Anyway it seems like MeadCo has already done what you intend to achieve with your ActiveX control. What matters would probably in how much better you control is.

And good luck with your project :)

Anonymous said...

Hey, I was cruising some links from iText when I found this article (great stuff, btw). I was pretty surprised to see my PDF stuff mentioned (the rbair package). So, here's the skinny -- the packages in the incubator have been shuffled about, but you should be able to still use the PDF viewer code I wrote. Its true that Sun has a pretty good PDF viewer, but that is different from the one in the rbair packages. My PDF viewer I wrote before coming to Sun, so they don't really have any say on that code in any case.

As for the Sun PDF viewer, I'm not sure what the status is.

My preferred work-around, BTW, is to simply use the JDIC embedded web browser to display the PDF files, leveraging the acrobat plugin and the fact that nearly everybody on the planet (in business circles) has the Acrobat plugin and a web browser installed on their system. Due to heavyweight/lightweight issues between Swing/AWT, you have to do a little trickery here, but expect that to be ironed out in upcoming releases of Swing.

Richard

Unknown said...

Hi Richard,
Thanks for your feedback. I did notice the rbair package in the java.net repository. It was present somewhere in the "attic". Thanks for reminding me to update my article on that.

Anonymous said...

If Your software is running on any UNIX-type OS , than I suggest using CUPS (Common Unix Printing System). It can print PDF files directly :). Try running "lpr" process from JAVA -works fine in my case. ("lpr -P printer pdfFile"). Soon CUPS will be also available on Windows platform :D

Anonymous said...

hi..i just hope my question is not off topic. anyway, the clients requested that whenever any type of files are uploaded, the system will automatically convert the files to PDF. I've been looking on solutions on how to this in java but couldn't find any. Creating PDF from scratch i would normally use iText. Manuall conversion of documents; I can use the printer drivers but auto convert?
Any advice? Thanks i advance.

Unknown said...

Hi Jade,
A possible solution to your problem might require you to read the contents of the uploaded files and then create the PDFs using iText from the content that your code has just read.
Alternatively, if your uploaded files are of specific types like images/movies etc., you can attempt to embed these files in a PDF document (again with iText).
I'm not aware of any library that would automatically convert input file streams into PDF documents.
Hope this helps.

Anonymous said...

why not something like


pdftops file.pdf - | lp

?

Unknown said...

Because your client may not want pdftops installed unless it becomes a well known tool?
Or because your customers maybe interested in paid-for-support despite getting something free?

I've been in situations where open-source software and libraries are a strict no, for trivial or non-obvious reasons, even though a cost-effective solution is achievable using OSS.

Anonymous said...

GHOSTSCRIPT IS THE KEY, WINDOWS OR LINUX:

gswin32c.exe -sDEVICE=png16m -dNOPAUSE -dBATCH -dFIXEDMEDIA -dDOINTERPOLATE -r150 -sOutputFile=c:\out.png c:\in.pdf

Run this command to create a temporary png version of the pdf and then you can use Java's Printing API to print the pdf programatically (able to select paper tray, copies, etc). OR, pay $500 + to one of the big pdf companies so they can sell you a product that may or may not give you results this good.

Kenny Grimm

Unknown said...

Hi,

Is there any way to print pdf programtically by giving the size, printername and number of copies?

Regards

Unknown said...

@Suresh : http://www.adobe.com/devnet/acrobat/pdfs/iac_developer_guide.pdf

A said...

Continuing on this fruitful discussion, is it possible to programmatcaly select the printer, page size and paper tray based on the filename of PS/PCL/PDF file...Yes I dont have PDF format limitation.

I am using FOP in Srping/Struts environment and IE 6+ browser on client side...and my lazy customer wants printouts on a particular tray with particular page size based on form FileName...after processing things frm the server.

I've tried using this snappy and flexible PrintFile tool but it doesn't provide advanced option like paper-tray etc.

Thanks.

P.S: Vineet your last response to Suresh was a thin hope for me on the same but didn't help much :(.

Unknown said...

You could try using pdfp obtainable at http://www.esnips.com/web/PDFTools

A disclaimer though, I haven't used it personally, so this advice is free.

Anonymous said...

Lots of great reading here, thanks! I had been seeking on yahoo when I uncovered your post, I’m going to add your feed to Google Reader, I look forward to much more from you.