Monthly Archives: October 2012

Format registry browser on Github

I’ve put the format-reg-browser project up on Github, in case anyone wants to play with the code. This is the first time I’ve committed code to any kind of Git site, but it looks as if the code’s really there. Let me know if there are any problems.

JHOVE 1.8b2

Oops… The Java 7 compiler on Ubuntu won’t build backwards-compatible classes, so JHOVE 1.8b1 wouldn’t run on earlier versions of Java. JHOVE 1.8b2 should fix the problem.

“Just solve the problem”

Running concurrently with National Novel Writing Month (aka NaNoWriMo) is “Just Solve the Problem,” an effort to get lots of people to attack the “formats problem” for 30 days.

Here’s “the problem,” slightly expurgated to avoid triggering nannyware:

In the last couple centuries, we’ve created a number of self-encapsulated data sets, or “files”. Be they letters, programs, tapes, stamped foil, piano rolls, you name it. And while many of those data sets are self- evident, a ****-ton are not. They’re obscure. They’re weird. And worst of all, many of them are the vital link to scores of historical information.

First thought: That’s not a statement of a solvable problem. It’s a statement of a situation which gives rise to many different problems. Still, throwing in some of my efforts can lead to professional contacts and maybe even a paying contract, and it’s the kind of thing I’d be doing anyway, so I’ve signed up for the wiki.

Extra points to anyone who can write a novel about the formats problem in 30 days.

Online file ID hackathon

CURATEcamp and Open Planets Foundation will hold a 24-hour (possibly more, due to time zones) online hackathon on file identification on Friday, November 16. The announcement says:

24hour+ live hackathon event where multi-time zone teams work on common technical projects related to the CURATEcamp iPres 2012 file id discussions.

Project proposals can be made by anyone.

We will start the day with New Zealand (GMT +12:00) and end with North America West Coast wrapping up project(s), hopefully with one or two solid deliverables by 12 midnight-ish PST (GMT -8:00).

JHOVE 1.8 beta

A beta version of JHOVE 1.8 is now available for testing. Please report any problems. New stuff:

  • If JHOVE doesn’t find a configuration file, it creates a default one.
  • Generics widely added to clean up the code.
  • Several errors in checking for PDF-A compliance were corrected. Aside from
    fixing some outright bugs, the Contents key for non-text Annotations is
    no longer checked, as its presence is only recommended and not required.
  • Improved code by Hökan Svenson is now used for finding the trailer.
  • TIFF tag 700 (XMP) now accepts field type 7 (UNDEFINED) as well as 1
    (BYTE), on the basis of Adobe’s XMP spec, part 3.
  • If compression scheme 6 is used in a file, an InfoMessage will report
    that the file uses deprecated compression.
  • In WAVE files the Originator Reference property, found in the Broadcast Wave Extension
    (BEXT) chunk, is now reported.

PDF/A-3

The latest version of PDF/A, a subset of PDF suitable for long-term archiving, is now available as ISO standard 19005-3:2012. According to the PDF/A Association Newsletter, “there is only one new feature with PDF/A-3, namely that any source format can be embedded in a PDF/A file.”

This strikes me as a really bad idea. The whole point of PDF/A is to restrict content to a known, self-contained set of options. The new version provides a back door that allows literally anything. The intent, according to the article, is to let archivists save documents in their original format as well as their PDF representation. Certainly saving the originals is a good archiving practice, but it should be done in an archival package, not in a PDF format designed for archiving.

Mission creep afflicts projects of all kinds, and this is a case in point.

A field guide to “plain text”

In some ways, plain text is the best preservation format. It’s simple and easily identified. It’s resilient when damaged; if a file is half corrupted, the other half is still readable. There’s just the little problem: What exactly is plain text?

ASCII is OK for English, if you don’t have any accented words, typographic quotes, or fancy punctuation. It doesn’t work very well for any other language. It even has problems outside the US, such as the lack of a pound sterling symbol; there’s a reason some people prefer the name US-ASCII. You’ll often find that supposed “ASCII” text has characters outside the 7-bit range, just enough of them to throw you off. Once this happens, it can be very hard to tell what encoding you’ve got.

Even if text looks like ASCII and doesn’t have any high bits set, it could be one of the other encodings of the ISO 646 family. These haven’t been used much since ISO 8859 came out in the late eighties, but you can still run into old text documents that use it. Since all the members of the family are seven-bit code and differ from ASCII in just a few characters, it’s easy to mistake, say, a French ISO-646 file for ASCII and turn all the accented e’s into curly braces. (I won’t get into prehistoric codes like EBCDIC, which at least can’t be mistaken for anything else.)

The ISO 8859 encodings have the same problem, pushed to the 8-bit level. If you’re in the US or western Europe and come upon 8-bit text which doesn’t work as UTF-8, you’re likely to assume it’s ISO 8859-1, aka Latin-1. There are, however, over a dozen variants of 8859. Some are very different in codes above 127, but some have only a few differences. ISO 8859-9 (Latin-5 or “Turkish Latin-1”) and ISO 8859-15 (Latin-9) are very similar. Microsoft added to the confusion with the Windows 1252 encoding, which turns some control codes in Latin-1 into printing characters. It used to be common to claim 1252 was an ANSI standard, even though it never was.

UTF-8, even without a byte order mark (BOM), has a good chance of being recognized without a lot of false positives; if a text file has characters with the high bit set and an attempt to decode it as UTF-8 doesn’t result in errors, it most likely is UTF-8. (I’m not discussing UTF-16 and 32 here because they don’t look at all ASCII-like.) Even so, some ISO 8859 files can look like good UTF-8 and vice versa.

So plain text is really simple — or maybe not.

Unicode

Words: Gary McGath, Copyright 2003
Music: Shel Silverstein, “The Unicorn”

A long time ago, on the old machines,
There were more kinds of characters than you’ve ever seen.
Nobody could tell just which set they had to load,
They wished that somehow they could have one kind of code.

   There was US-ASCII, simplified Chinese,
   Arabic and Hebrew and Vietnamese,
   And Latin-1 and Latin-2, but don’t feel snowed;
   We’ll put them all together into Unicode.

The users saw this Babel and it made them blue,
So a big consortium said, “This is what we’ll do:
We will take this pile of sets and give each one its place,
Using sixteen bits or thirty-two, we’ve lots of space

   For the US-ASCII, simplified Chinese,
   Arabic and Hebrew and Vietnamese,
   And Latin-1 and Latin-2, we’ll let them load
   In a big set of characters called Unicode.

The Klingons arrived when they heard the call,
And they saw the sets of characters, both big and small.
They said to the consortium, “Here’s what we want:
Just a little bit of space for the Klingon font.”

   “You’ve got US-ASCII, simplified Chinese,
   Arabic and Hebrew and Vietnamese,
   And Latin-1 and Latin-2, but we’ll explode
   You if you don’t put Klingon characters in Unicode.”

The Unicode Consortium just shook their heads,
Though the looks that they were getting caused a sense of dread.
“The set that we’ve assembled is for use on Earth,
And a foreign planet is the Klingons’ place of birth.”

   We’ve got US-ASCII, simplified Chinese,
   Arabic and Hebrew and Vietnamese,
   And Latin-1 and Latin-2, but you can’t goad
   Us into putting Klingon characters in Unicode.

The Klingons grew as angry as a minotaur;
They went back to their spaceship and declared a war.
Three hundred years ago this happened, but they say
That’s why the Klingons still despise the Earth today.

   We’ve got US-ASCII, simplified Chinese,
   Tellarite and Vulcan and Vietnamese,
   And Latin-1 and Latin-2, but we’ll be blowed
   If we’ll put the Klingon language into Unicode.

JHOVE format notes

New on my business website: JHOVE format notes.

Preservation in the geek mainstream

Digital preservation issues are gaining notice in the geek mainstream, the large body of people who are computer-savvy but don’t live in the library-archive niche. Today we have an article in The Register, “British library tracks rise and fall of file formats.” It cites the British Library’s Andy Jackson, supporting the view that file formats remain usable for many years, even if they’re no longer the latest thing.

The Register article is short but nicely done. It naturally skips over issues which Andy’s original article deals with, like just how you reliably determine the formats of files. What’s significant is that it shows that concern about the long-term usability of files isn’t just a concern of a few specialists.