MeatballWiki | RecentChanges | Random Page | Indices | Categories

Notes for and response to JonUdell?'s [paper] on GroupWare, Internet Groupware for Scientific Collaboration follows.

Critical scientific behaviour

In the introduction he claims that ideal net.groupware will be



but in the body of the article he contradicts himself when he actually claims (not necessarily overtly) ideal groupware should


Event coordination

Of the three examples cited, only evite.com still exists, even the Yahoo! invite service is closed. Evite is also now grossly outmatched by meetup.com (cf. MeetUp?), which was heavily used during the HowardDean campaign.

Nonetheless, Jon envisions event coordination web services to act like the Outlook meeting schedule, except more friendlier. It organizes the invitations, organizes the attendees' time, and then helps elect a suitable time and place. One drawback of these systems is the painful schedule negotiation phase, which quickly becomes tedious if you have multiple meetings to schedule all at once. Manually inputting your availability makes the services not worth the effort.

MeetUp? is not nearly as powerful as these services, but perhaps it succeeds since it doesn't try to do schedule negotiation.

Group discussion

Mailing lists are the most common form of group discussion on the Internet, however they are structurally broken because they decentralize the actual places of conversation to individual mailboxes rather than bringing everyone to one place. While there may be web archives available, they are most ignored since the actual interface for discussion still is mostly personal inboxes.

Moreover, e-mail isn't designed to be structured; it's meant to be single, unidirection, and point-wise. This makes tracking the thread of discussion on a mailing list very difficult, compounded by varying compliance to the e-mail standard. Although e-mail now contains information about reply relationships, not all mailers follow this. Thus, we use the subject line to connect related mail together, even if this makes the actual content of the discussion opaque to someone just seeing fifty messages with the same subject line.

Further, e-mail isn't secured or multimedial, although the web supposedly makes this better. Actually, e-mail is multimedial, but most of that is spam.

In the olden days, UseNet (~NNTP) used to be very useful for group discussions, but Usenet is mildly dead now. Jon waxes on about a better UseNet (that looks suspiciously like BulletinBoardSystems), but we'll just ignore that like the Internet has ignored every other such idea. For those wondering, usemod.com, the site that hosts Meatball, stands for "Usenet moderation", another attempt to save UseNet.

His discussion of why NNTP is better than e-mail revolves around the centralization of NNTP in one space, which can be administrated by one person (e.g. security), although his strawman that e-mail is only plaintext gets to be a bit worn by this point since Usenet spam is just as, um, colourful.

He describes two mailing list services as well. QuickTopic? creates a mini-discussion board with a random name so that it can only realistically be found from where you want to originate it from. This keeps the discussions small, focused, and "disposable." QuickTopic? is very popular.

Ka-Ping Yee's Roundup is offline, but the idea was to help bug and task tracking. Just forward an e-mail from a list discussion to Roundup, and it creates an item around which further discussion may flow (a mini-mailing list). Joining the list is simply a matter of having your e-mail address cross Roundup's path vis a vis that item. Unsubscribing is unnecessary (in theory) since discussions are so narrow they will die down quickly, e.g. after a bug is fixed. (Roundup sounds like a spam cannon to me.)

He also mentions wikis, which are designed (in part) to fix a lot of these problems with mailing lists. For one, threaded discussions get reworked into document mode frequently. (cf. ReworkingThreadMess) Plus, you can avoid getting into threads in the first place by just writing directly into the document you want to fix. Further, discussions are centralized in one place (the wiki), and organized by page name. Yet, wikis structure is not well understood by a lot of people, so they won't serve to replace mailing lists.

Broadcasting and monitoring news

What Jon talks about here, really, is keeping up with things in the face of InformationOverload. He mentions RichSiteSummary, which is a terrible standard. XML nazi, I cringe when he says that RSS is touted as a successful application of XML, since it is just so ugly. It's laughable that Jon suggests you can create such things by hand. You don't even want to see MeatballWiki's output. It makes me cry.

Mark Pilgrim takes a big [swing] at its incompatibility, and Sam Ruby's [Atom] project is attempting to remake it in the image of sanity. RSS is reaction against the massive decentralization of weblogs, something that Jon in his own article has arguments against.

If you're intrigued, however, in Jon's "pushbutton webpublishing" idea, you would enjoy Joshua Schacter's wildly popular http://del.icio.us, as it is a no-nonsense take at weblogging. You visit a site, and you post its URL with a brief description and some keywords if you wish, and that's about it. The system then adds some structure to help you discover other people who are interested in the same things you are, and then you can peel off their links too. But there is no pushback, no attempt to make relationships with other people. Just the links and that's it, despite pressure to add a wiki to their backend (which I have tried to argue against doing, successfully for the moment).

Scientific Publishing

TeX?, Word, and FrameMaker? still dominate the publishing side. Jon laments at the crappiness that differing file formats makes life, which is true. It's very useful to cite a particular sentence in a document directly, or an image, or to slice out a chunk of text or an image and "quote" it through a TransClusion. More to the point, there are no web-native formats for scientific writing yet, although MathML? and SVG are coming along alright. Since I worked with the W3C SVG committee, I won't comment on what I think of what SVG sekretly is, but I definitely like it even though it presents a front of ridiculous over-complication.

-- SunirShah


MeatballWiki | RecentChanges | Random Page | Indices | Categories
Edit text of this page | View other revisions