File formats for information exchange

[mdlbear]

This blog post by Eugene Spafford points out that

Microsoft Word:

1. Is not a document interchange format -- it is not designed for document transport
2. Is not installed on everyone's machine, nor available for everyone's machine
3. Not all versions of Word are compatible with each other
4. Results in huge, bloated, files for tiny content (such as memos)
5. And of course, Word is commonly a vector of viruses and maicious hacks.

He includes a link to the (plaintext) "bounce message" that he uses to educate people who send him Word documents. Well worth a look.

Similar considerations apply to files produced by PowerPoint, Finale, Photoshop, and other programs. Open-source, cross-platform programs like Audacity and the Gimp aren't immune either: even when they're available cross-platform, you shouldn't use them for email.

If you're actively collaborating with somebody who you know is using the same program and version (I'm upgrading to Audacity 1.3; this isn't just a Microsoft problem by any means), by all means use that program's file format, but put it on a website and email the URL rather than trying to ship the whole darned thing in email. Many email systems will bounce big files anyway.

If you're sending a finished product, use plain text, HTML, or XML if at all possible. Other text-based standards include LaTeX for typeset documents and ABC for music. Images can be sent as PDF or JPEG, formatted documents as PDF, music notation as a zipped MIDI file, and sound files as FLAC or Ogg. But even here they're big enough that you'll want to put them on the web and email a link.

(via spaf_cerias.) ETA Oh, and if you're thinking this was inspired by a particular piece of email, or a particular blog post, it wasn't. Just seemed like a good reminder, especially now that I'm trying to put together a CDROM.

Comments

[andyheninger.livejournal.com]

I hate MS Word attachments too.

But then, I dislike mail attachments in general. I much prefer getting mails with URLs referring to the documents on the web. Maybe I really didn't want to download that multi-megabyte whatever.

Moving to file formats, I also have some opinions.

Plain Text. Not so good. Will the recipients consider it to be Latin-1 or Windows 1252 or Shift-JIS or UTF-8 or ...? Even without charset problems, there's still no telling how the text will be displayed. Fixed or variable pitch font? Line wrap? What becomes of tabs? There's no way to do even the most primitive of formatting with any assurance of tolerable results. Even if the source started out plain text, distribute it as HTML with a correct charset declaration and enough formatting markup to get it to display sanely.

PostScript. Nothing comes preinstalled on Windows to display or print it. GhostScript for Windows is big, confusing to install and generally annoying. You can't copy the text. Stick with PDFs.

Movies. Movies I've been having trouble with. I have been posting videos of Maddie's band for the band families, and haven't found a good solution yet. Windows Media Format displays decently on both Macs and Windows, but it's yet another MS format, and not all Mac owners have consumed the Microsoft coolaid, er, installed the Windows Media stuff. QuickTime works fine for Macs, but QuickTime for Windows requires a faster machine than many people have to display with decent quality (640 x 480 x 30FPS, retaining reasonable detail). I haven't tried playing either format on Linux.

It'll be interesting to see where things are in a few years. Maybe ODF will come on strong.

[mdlbear]

Plain text: the character set can be indicated using MIME; there used to be conventions for formatting email (8-column tabs, 72-column lines, fixed pitch), but most email users these days are probably unaware of them.

HTML: I agree that that's the way to go if formatting is important and the text was created in any kind of word processor. Put it on a web page and hand me the URL in email; as far as I can tell the vast majority of HTML email is spam, so I tend not to read it unless I have good reason to think it's OK.

PostScript: Both PS and PDF are potentially dangerous, and annoying. Hopefully some combination of ODF for editable documents and real image formats like JPM and SVG for page images will help.