Author Topic: Unorthodox Behaviour (not the record by BrandX...)  (Read 1266 times)


  • Newbie
  • *
  • Posts: 1
  • Karma: 0
  • Dinosaur programmer
    • No programming here :)
Unorthodox Behaviour (not the record by BrandX...)
« on: May 31, 2014, 04:39:40 pm »
I've been trying to generate simple separator pages (Item 1, Item 2 etc) on, in essence, A5 format. Initial try with M$ Word, but after numerous tries to get the spacing correct when printing two pages on one A4, I gave up and moved to Scribus. I had used the program once before to create little forms, that were subsequently printed and cut by a professional print-shop.

First problem I encountered was the fact that Scribus has no option to add sequential numbers to items - abusing the page numbers as suggested in other posting on this forum is not an option, as that would bring me back to the problem of getting the print properly aligned when printing two pages on one page.

So I did what I've done numerous times with RTF files coming out of M$ Word, I had a look at the generated XML. The first thing I did was removing all empty attributes, i.e. 'value=""', and any numerical attributes that are 'value="0"', rightly assuming that these would be set to their defaults when missing.

What I did notice were the, sorry if this is a (too) strong opinion from a newbie (who's been working in IT on IBM mainframes since 1985...), ludicrous high-precision values used for things like size and position. Are we positioning items with positions accurate to the nearest micron?  :o

If anyone would like to print a 4A0 sized sheet of paper (1,682 x 2,378 mm / 66.2 x 93.6 in) at 2,400dpi, a fixed point 18-bit number would be more than adequate to describe placement of items, but a 32-bit word split in 24:8 fixed point format would probably be a good compromise between precision and speed.

But back to the subject...

Having created a two page .sla file with my 2-up separator pages, the next thing was to just add <PAGE ...> and <PAGEOBJECT> items, and there I stumbled upon another user-unfriendly (for those users that edit the .sla file directly) feature, positioning is not on a page-by-page basis, but it's based on absolute positions from the very start of the document, and using these high-precision values the calculation of the position of a next item becomes a bit harder, but to solve that, I just rounded the 15 digit values to something that makes a bit more sense, and lo-and-behold, even with five pages, the placements are still accurate to 0.001 mm (aka one micron), not very surprising, as the unit used in Scribus are points, and there are already 2.834645669 points in one mm - it's likely that my rounding to 6 decimal places is still way-over-the-top, 4 would quite likely be more than enough.

So, I'd like to end by suggesting the developers of this fine piece of software (no pun intended, it is really a fine piece of software) that the .SLA format could do with a bit of tuning, default attributes shouldn't be written out, 15 digit numbers are way over the top (and indicate that the program works with floating point numbers) and positioning should be relative to the current page. An option, already asked for in these fora, to for auto-generated sequential numbers would also be very nice.

For those interested, I have attached the original .sla and my hand-optimized version in the attached
Robert AH Prins
robert.ah.prins @ the.15+Gb.Google thingy
No programming here :)