New DriveThru G2A

[PhilHibbs]

I just downloaded the updated "Glorantha The Second Age Core Rulebook" from DriveThru, and the colour map of Glorantha has gone from the second page. Why was this removed?
*Edit* Also, some of the fonts have been changed from the stylish serif font to Arial, and the title "Welcome to Glorantha" looks awful with characters overlapping.

 

[sdavies2720]

That looks like it's substituting one font for another when it renders -- maybe the original font is no longer embedded in the file and it's not on your system? So it's looking for the 'substitute' one. If the header were specified as Arial, the spacing would probably be ok.

Unfortunately I don't have that file (or know what's up with it specifically)

Steve

 

[languagegeek]

I just downloaded the updated "Glorantha The Second Age Core Rulebook" from DriveThru, and the colour map of Glorantha has gone from the second page. Why was this removed?
*Edit* Also, some of the fonts have been changed from the stylish serif font to Arial, and the title "Welcome to Glorantha" looks awful with characters overlapping.

It looks to me like when the PDF was exported, "embed fonts" was not selected. Not only is the title font gone, but the text font is different as well. The original is in Adobe Garamond and has nice italics. The PDF version you show has an ugly computer-generated slanted hack Times New Roman.

That sans-serif is something your computer seems to have grabbed at random, even Arial isn't that bad.

 

[simonh]

The substitute font looks like it's kerned (inter-character spacing) for the original 'scripty' font. The substitution has to be done that way to make sure the substitute font fits into the same space as the original.

It does look like the font wasn't embedded in the document. Whoever generated the pdf wouldn't know because they'd have the font installe don their system. I get the same problem in software development sometimes. You can try to package up all the library dependencies into a distributable version of an app, but it's a pain to test as you can never be certain the boxes you test it on are really 'clean'.

Simon Hibbs

 

[languagegeek]

While we're on the topic of the PDFs. I wonder how the typesetter at Mongoose handled "fl" "fi" ligatures. I've got the Corebook PDF, and if I do a search for "fly", it only finds instances where the "f" is capitalised: Fly. The lowercase f's don't show up. If I do a search for "official", I get nothing.

It looks like there was some search/replacing done for "f+l" to "fl" (Unicode U+FB02) and "f+i" to "fi" (U+FB01). Either that or the version of Garamond you're using has bad character names (unlikely). I assume you're using Adobe inDesign or Quark, so turning on "ligatures" will give nicely shaped fl's but still remain searchable.

 

[languagegeek]

Whoever generated the pdf wouldn't know because they'd have the font installe don their system. I get the same problem in software development sometimes. You can try to package up all the library dependencies into a distributable version of an app, but it's a pain to test as you can never be certain the boxes you test it on are really 'clean'.

That's why I always keep an old vanilla XP laptop kicking around. If the document looks good on that, it'll look good on anything.

 

[PhilHibbs]

It looks like there was some search/replacing done for "f+l" to "fl" (Unicode U+FB02) and "f+i" to "fi" (U+FB01). Either that or the version of Garamond you're using has bad character names (unlikely). I assume you're using Adobe inDesign or Quark, so turning on "ligatures" will give nicely shaped fl's but still remain searchable.

So it can create a PDF that uses the ligatures for display but has an "alternative" searchable version of the text as well? Actually, I think I've seen scanned pdfs that are searchable - what you see is the scanned bitmap, but there's a text representation hidden somewhere that the search uses.

 

[MongooseMatt]

Hi guys,

We are looking into all of this!

 

[PhilHibbs]

We are looking into all of this!

It hasn't affected me negatively, since the name of the pdf file has also changed so I still have the old version, and I did check the new file in my Downloads folder against the old one and noticed it was half the size so that rung alarm bells. DriveThru also presented Races and the core RQ2 rules as "changed since your last download" files, and I've downloaded them and compared them against the copies I already have and haven't noticed anything different, but I haven't checked every page.

 

[languagegeek]

It looks like there was some search/replacing done for "f+l" to "fl" (Unicode U+FB02) and "f+i" to "fi" (U+FB01). Either that or the version of Garamond you're using has bad character names (unlikely). I assume you're using Adobe inDesign or Quark, so turning on "ligatures" will give nicely shaped fl's but still remain searchable.

So it can create a PDF that uses the ligatures for display but has an "alternative" searchable version of the text as well? Actually, I think I've seen scanned pdfs that are searchable - what you see is the scanned bitmap, but there's a text representation hidden somewhere that the search uses.

In a nutshell. There are characters (underlying), and there are glyphs (surface). Each character has a Unicode index number. That index number is consistent from computer to computer - i.e. "à" on a Mac is the same as an "à" on your PC. It didn't use to be that way, but Unicode is now virtually universal.

So, "f" has a unicode number (0066), "l" is (006C) and "y" is (0079). When I do a search for "fly", I type those letters in the search box, and the computer looks through the PDF for 0066+006C+0079.

However, proper typography requires ligatures between certain combinations of letters, the most common being "fl" "ff" "fi". In the old days, the solution was to make up unique new characters for these ligatures. When Unicode came along, these precombined characters had to be accounted for for compatibility purposes, so "fl" was given the code (FB02) and so on. Typographers using old fonts or old software continued the tradition of replacing every "f+l" (0066+006C) with (FB02). Looks fine on the printed page, but in the digital age, there are problems because if I want to search for "fly", I won't find anything because the "f+l" has been replaced with "fl" (FB02).

What good-quality modern fonts and half-decent desktop publishing software do is this: every time there is the character sequence "f+l", use a surface glyph ligature "fl". The glyph does *not* alter at all the underlying characters, the change is purely cosmetic, superficial. The underlying sequence remains "f+l". Thus my search for "fly" will work perfectly. In this day and age, the legacy precombined ligatures like "fl" (FB02) should not be used.

Searchable scanned pdf's are a different story. Here someone did a OCR run of the scans and makes a background script to accompany the scanned text. These scripts aren't perfect, and really run into trouble when non-English words are present.