Reconfiguring !Printers

I often have to send engineering drawings to my clients in PDF format, usually created on A3 paper size in !Draw. For many years I have been using the Postscript printer driver and !RiScript to do the conversion to PDF. This has worked very well for me over the years except for one thing…..
Virtually everyone uses a landscape monitor but an A3/A4 laser printer changes its printing orientation depending upon whether A3 or A4 paper is in use. Because I am dependent upon the !Printers application, A3 documents, when converted to Postscript with !RiScript, show in portrait mode, even when they originate in landscape. When I send the PDF onwards for display on a Windows machine, it is an easy job to rotate the drawing in Adobe Acrobat. However, I seem unable to prevent this from happening in the first place.
Just recently, one of my recipients has asked “Why do I have to keep turning your documents round in order to see them on my screen?” With large resolution screens being more commonplace these days, there is less need to print out drawings and so this is more of a problem.
I guess that the problem goes back to the fact that the printer driver fills the page from the short-side first when operating in A3 and this needs to be changed.
How can this be done please?

I don’t know if it will fix your problem but PrintPDF here (Which uses Ghostscript) is probably the best way to create PDFs!
The only downside is Ghostscript needs a long final names format drive which is unlikely to effect many people.

Hi Chris,
I’ve done some experiments with Ghostscript and PrintPDF and found it not to be as dimensionally accurate as using RiScript which is why I have been using it for so long.

I’ve done some experiments with Ghostscript and PrintPDF and found it not to be as dimensionally accurate

You may want to include the PS2Paper utility on the same site in your testing too. Paper orientation and dimensions are interlinked if you think about it.

I have just checked some landscape documents created in our PDFsuite (aka PDFmaker) software, and can confirm that they are shown in the correct alignment on a Windows system. For example, the SafeStore2 manual is included on its CD as landscape A4 with two “A5” pages per A4 sheet. This appears correctly presented on my Win10 PC using the built in PDF viewer.

I was doing other testing yesterday for a musician customer with irregular page sizes for music printing/publishing, and those also seemed fine too.

Of course, there may be issues with how the originating software orients the page (eg. printing portrait with rotated contents vs printing landscape with non-rotated contents).

But, certainly, it shouldn’t be difficult to achieve your desired results. I know that PDFmaker has been used for a variety of professional print purposes and online publishing services. I’m sure PrintPDF likewise. I suspect that “dimensional accuracy” is needed in most such scenarios, otherwise books would fail to be printed correctly, or music badly laid out? To be honest, I’m not 100% clear what you mean by “dimensionally accurate” in this sense, but I think what you’re trying should be perfectly possible.

Further thought – one scenario where things do tend to go wrong is if a document mixes alignment/sizes for different chapters. Ghostscript tends to align everything based on the first page of the job (if memory serves – it is a good 3+ years since I looked into this). The solution is to print such chapters separately and join them together. This should correctly orient each chapter when viewed in a PDF viewer. I think (fuzzy memory).

I’ve done some experiments with Ghostscript and PrintPDF and found it not to be as dimensionally accurate as using RiScript

What do you mean by this? PrintPDF just calls Ghostscript for the hard stuff (as does PDFMaker), and I’d be surprised if it got dimensions wrong. As Steve says, paper sizes might not be what you expect when using the Level 2 PostScript drivers, but otherwise it should work fine. I’ve certainly used the system for creating some dimensionally important stuff (inserts for CD jewel cases, other copy for accurately printing and guillotining), and it’s always seemed fine.

As far as I know, Ghostscript isn’t told the orientation of the page, and guesses based on the content sent to it. As with Andrew this is a fuzzy memory, but since I think GS uses lines of text to do its guessing, an engineering drawing which uses a vector font for text might not have any clues as to which way the page was oriented. I would guess this is also true of RiScript, as I’m not sure that the RISC OS printing system even knows the content orientation (or passes the info on if it does).

(removed (not relevant))

Because I am dependent upon the !Printers application, A3 documents, when converted to Postscript with !RiScript, show in portrait mode, even when they originate in landscape.

Okay, I’ve just tried this. I am using standard Draw to create an A3 sized PDF.

I have created a paper size called A3 with 5mm margins around (created using the drivers themselves, not the PS2Paper utility, though that looks useful so I’ve grabbed a copy, thanks). The reason for this is that it will appear with a small grey outline in Draw with Misc → Paper limits → Show ticked. This is so I can verify that the right paper size is being invoked, and that it is obeying the landscape setting.
It is.

So I print. The printer is PostScript2 to file as PrintPDF expects. Sorry, I have not used RiScript in ages because I found its rendering to be highly eccentric. It is probably better these days, but being a commercial product, I don’t have a copy to work with. So, it’s PrintPDF/GhostScript.

If it matters, Printers v1.83 2014/02/19. PrintPDF 0.88 2012/08/04. And Draw 1.30 (2016/10/27) with a self-built low-vector ROM dated 2017/01/24.

The result is here: http://heyrick.co.uk/random/a3landscape.pdf

Displays as expected (landscape) in PDF (3.02.1.24¹ 2011/07/11) under RISC OS, Adobe Reader (version unknown?) under Android, and Adobe Reader XI on Windows (was so damn slow I killed the task before I got to Help → About to read the version number).

Without RiScript to test, all I can tell you is that with PrintPDF, my simple test file looks correct.

Are you able to share a DrawFile that may be tested?

With large resolution screens being more commonplace these days, there is less need to print out drawings and so this is more of a problem.

Yes, it makes sense with displays more often being widescreen to have the document fit sensibly.

I guess that the problem goes back to the fact that the printer driver fills the page from the short-side first when operating in A3 and this needs to be changed.

Huh? PostScript is a page description language. The printer itself will ‘draw’ the document to an internal bitmap and then print it off as required. The printer driver should use a method not unlike a Wimp redraw loop to get the data for the printing, and for what it is worth, Draw objects are generally directly translated to PostScript objects (rather than via a bitmap). See https://www.riscosopen.org/viewer/view/castle/RiscOS/Sources/Printing/Modules/PDModules/s/PDriverPS/Draw?rev=4.2;content-type=text%2Fplain…

I notice that the PostScript generated by Printers, besides being rather obscure, does not explicitly specify an “orientation”, it seems to define the page dimensions as “it’s an A3 page”. Hmmm.

Okay, delving into it a bit further. Comparing A3 landscape with A3 portrait, the differences are:

47000 MP 1162581 MP T
0 -65536 65536 0 UM

vs

28118 MP 470000 MP T
65536 0 0 65536 UM

It looks as if MP is 400 div, T is translate, UM is a lot of 65536 div and roll. Can’t be bothered to try to find my old PostScript reference to work out what it is doing. ;-)

Looks, at a purely half-assed guess, that it is defining the page orientation by using a translation onto a portrait A3 page rather than just saying “the paper is sideways”. Hmmm… I guess GhostScript is clever enough to guess that this means the page is landscape rather than the literal interpretation of “print this stuff sideways”.

In which case, GhostScript is doing what is logically correct, while it may be that RiScript is doing exactly what it was told…

How can this be done please?

Can we start with a PostScript driver that outputs PostScript that is actually readable? The stuff in !Printers.ps.PSFiles.Level2.PSprolog2, and subsequent output generated with it as a library, is really horrible to try to decipher. :-(

¹ Seriously? Version three point oh two point one point twenty four? <facepalm>

Adobe Reader XI on Windows (was so damn slow I killed the task before I got to Help → About to read the version number).

A.R. – So much of a lumbering juggernaut of unwanted feature bloat that I ditched the pile of poo on PC at work and loaded Foxit Reader.

Can we start with a PostScript driver that outputs PostScript that is actually readable? The stuff in !Printers.ps.PSFiles.Level2.PSprolog2, and subsequent output generated with it as a library, is really horrible to try to decipher. :-(

What you really need is for someone to spend the time updating the Postscript driver to a (reasonably) current version.

Surely we have a PS3 driver – it’s just not free!

Surely we have a PS3 driver – it’s just not free!

Here you go folks, a cheap hardware platform with a free OS…
…but you have to shell out for the printer drivers that every other OS bundles in for free.

Or do all your lovely drawings in !Draw on your Pi, convert them to SVG and email them to your Mac, then print them from there, generally after importing them into a LibreOffice document because they’re part of a bigger document anyway.

That’s it, aim for the simplest method – even if most people think it’s way too much faff and rush off to platform that doesn’t give you that entertainment.

…but you have to shell out for the printer drivers that every other OS bundles in for free.

For me, it isn’t so much the cost element as the fact that it is expected standard behaviour these days.

Rather like a filesystem that can fix itself, pretty much standard since DOS 6.2 when SCANDISK arrived (now known as chkdsk), and fsck and variants on Linux machines.
RISC OS? We have *CheckMap that is only competent to tell you the map is good or bad. It doesn’t tell you where it is bad, nor have I ever witnessed it rewriting a bad map with a copy of a good map.
No, repairing anything requires a commercial utility that doesn’t appear to be updated much any more.

Here are the big three that may bite: “Map inconsistent with directory tree”, “Bad free space map”, and “Broken directory”. I would imagine the first two could be dealt with by walking the filesystem and rebuilding. The “Broken directory” could be resolved, perhaps, by trying to sanitise the directory contents. Doing this may leave orphan fragments (allocated space that isn’t attached to anything, but it’s better than just hitting a brick wall. Leave the fancy recovery to the commercial product, at least let’s have something to get the filesystem to be able to recover itself to a basic degree…

Or do all your lovely drawings in !Draw on your Pi,

Have you tried OpenVector? It is Draw with go-faster stripes. Lots of them.

If that looks like overkill, then I suggest you try DrawPlus. It’s like all the things Draw should have been, without being too unlike Draw.

http://www.users.on.net/~belles/software/openvector/

I did try DrawPlus long, long ago – bought a copy for the publications office at The Physiological Society, where I ran the pre-press operation for their journals (originally on A540s, later on Risc PCs). I think I only bought a single-user copy, with a view to getting a site licence if it looked useful, but plain !Draw did everything we needed okay. Since that was probably 1994 I expect DrawPlus has advanced a bit since then!

I’ve not tried OpenVector – thanks for alerting me to it. Will give it a shot. Not that I’ve any complaints about !Draw – everything I wanted that it didn’t do out of the box I’ve just written little apps in BASIC to provide for me, some of them available on my website, some still too scruffy to offer publicly.

No, repairing anything requires a commercial utility that doesn’t appear to be updated much any more.

It’s not as if you need to update a disc checker every week. DiscKnight works on all the new hardware, and was updated to cater with the new filecore with 4 GB file limit. What more could you ask for?

On the other hand, we could open a bounty to buy the rights to various stuff to be bundled free-of-charge with RISC OS. I nominate the PostScript 3 printer driver, SparkFS (full version), SystemDisc and DiscKnight.

It’s not as if you need to update a disc checker every week.

I was thinking of the impending change to a possibly partition-friendly filesystem. But maybe more than this, our only file system recovery tool is a closed source commercial product.
This isn’t to demean druck, it’s a good product. More a strong criticism of Acorn never bothering to have tools for data recovery in case of problems. Hell, Broken directory is older than the hills…

PS: Apologies for the typos in the original version of this message. I was writing in a hurry to send it before hitting the forest and losing the signal.

On the other hand, we could open a bounty to buy the rights to various stuff to be bundled free-of-charge with RISC OS. I nominate the PostScript 3 printer driver, SparkFS (full version), SystemDisc and DiscKnight.

Sounds like a reasonable idea. Of those you’d have to ask whether the authors actually make any money. SystemDisc and PS3 possibly, the other two probably barely cover the cost of a cup of coffee per month.
For SparkFS the free upgrades mean that most of us haven’t had to look at a cost for the fixes in the last decade or more. I’d say the same for DiscKnight but I can’t find the registration so I haven’t upgraded that in a decade.

Sounds like a reasonable idea.

I agree… completely.
But IMHO it should not be a bounty.
Best is to ask author for a price, then to ask users for money.

Best is to ask author for a price, then to ask users for money.

Err, isn’t that what is already being done? Author has a price, users pay it if they want the software….?

Correct :)

I’m gradually making the sources of my software freely available. I’ll get there with SparkFS. I do appreciate people who have bought it in recent years. I’m also happy to think that some people bought Spark (5.99) then ArcFS (5.99) and got SparkFS for the offer price of 12 quid in 1992 and its been free upgrades ever since. When I use archives on Windows and Linux I am impressed how much better SparkFS is (due to the support in RISC OS). There’s not a big win from a free SparkFS, because the read only version does most stuff.
Rick’s talking about corrupt discs above – I vote for a file system that does not corrupt discs. ‘cos yeah Windows fixes corrupt discs but it does not tell you what you’ve lost.

I’m also happy to think that some people bought Spark (5.99) then ArcFS (5.99) and got SparkFS for the offer price of 12 quid in 1992 and its been free upgrades ever since.

That long? If I dig around I may find the original disc.
Mind you given the extensive piling system in that room the dig could require a few months.