[Up] [Map] [Prior-Yale critique] [Robot Wisdom home page]
From: jorn@mcs.com (HyperTerrorist) Newsgroups: alt.hypertext,comp.infosystems.www.misc,alt.culture.wwwSubject: WWW-RANT: Hypermedia Joyce Studies, an online journal
Date: Thu Jan 04 16:06:04 CST 1996 Summary: a detailed critique of an academic WWWeb site[This rant is for educational purposes only. Not to be taken personally.]
HyperTerrorist has examined their unhappy first attempts at hypertext design guides-- the customarily narrow formalities of academic publications result in preposterously _inflexible hypertexts, that hang off the WWWeb like great splintery chunks of plywood...
And then, too, there's the pomos, who've been wanking on about hypertext theory for a decade already, but now have the opportunity (or, more likely, the burden) of laying out their genius for the WholeWideWorld to see...
So the first issue of a new journal, Hypermedia Joyce Studies is a ripe occasion for HyperTerrorist to don his hyper-togs, and set the WWWeb to rights!!!
Joyce studies are a favorite stompingground for the pomos. Joyce, in fact, is the colossus in whose shadow they've taken root ...because "Finnegans Wake" is a text so complex it seems a license to pure literary anarchy...
HyperTerrorist has his own views on this question. If Joyce claimed he could justify every letter of FW, then we ought to be wary of those who'd use it to rationalize hypertext chaos...
But there's a terrifying case of pure hypertext chaos at the center of HJS#1, designed, in fact, by one of the editors: Lawrence James.
One knows immediately, by the way the title contains two different embedded links, that this person is eager to clobber us with his hyper- daring. One doesn't know, without delving deeper, that this essay is broken into some dozen files, interlinked in what seems an entirely arbitrary fashion. [Note to RC: you can combine the 'name' feature and the 'href' feature into a single anchor. Lynx has display-problems with the doubled tag.]
Happily, there's no intelligible content anyway, so we can ignore this essay. The errors aren't even interesting enough to analyse...
A more educational attempt at hypertext design is offered by Michael Ditmore in his discussion of a Joyce-MOO
Ditmore's error is simply to be 'link drunk', breaking out paragraphs of his essay to be separate pages, so that his ~5-page article is spread out over some dozen WWWeb-pages. If the net were instantaneous, this might not be a problem, but under realistic conditions it becomes a sort of 'gotcha' water-torture...
The topology of these subpages is also confused. One link leads off to a chain of short pages, with no hint of the future direction, and (worse) no guidance for navigating back. (Each page has a lone HJS button at the top that leads to the journal's own title page, skipping all intervening levels.)
A striking peculiarity about the other essays is that they all handle footnotes differently! Footnotes have always been the prototype for why we must have hypertexts, but it's not at all clear how they should be implemented...
Do you make the anchor-text a note like "[1]" or "(Ellmann p366)" in imitation of academic conventions? Or do you highlight part of the text itself, as the WWWeb does?
Rice's essay tries the latter, but the effect is a sort of 'gotcha' where you're expecting a new page of text, but get only a footnote... so this strategy must be wrong. (How would you know you weren't missing text unless you tried every link! This is torture...)
Should each footnote get a page of its own? Surely not, and happily (even unexpectedly) none of the HJS essays attempts this, leaving them as a single body of text. But should they be in a separate file, or all part of the main textfile?
Theall opts for the former, Rice for the latter.
A separate file has the disadvantage, with browsers that lack caching, of having to be reloaded each time you jump from text to note and back, so HyperTerrorist recommends the former approach, including text and notes in a single file. It's possible, though, that the best thing to do is to integrate the notes into the text itself!
This is paradoxical, but links in WWWebspace are invitations to explore, where footnotes are deadends for deadheads. If the notes offer no added text, they might just be numbered and implemented without live links. ( Roughley adds a odd new quirk: he's got live "[1]"s in the text, but no corresponding numbers on the notes themselves!)
Notes that do offer side comments fall under the same rules for page length as other hyperlinks-- if it's only a paragraph or two, you probably ought to just include it with the text-- the tradeoff between lagtime and info-gained is just too feeble otherwise.
The central principle of hypertext design is not offering opportunities for exploration but rather offering opportunities to ignore inessentials.
This requires that all links should offer enough of a summary of their targets that the reader can choose intelligently whether or not to follow the link. By this model, tables of contents should resemble the 'analytic ToCs' of British teach-yourself texts, summarizing the contents of each chapter in the most concise way possible, so that one can quickly see whether it offers something she needs...
(The trend in academia, unhappily, is to withhold clear summaries, so as to exaggerate the intellectual stature of the author. One may hope the WWWeb will work against this trend.)
Links to GIFs-and-nothing-else should declare this, and even mention their size. It's nice to distinguish offsite from onsite links as well, since they imply a different loading-delay ("latency") and a different sort of return-path as well.
There's a spelling error ("inagural") on the first page.
The journal offers a simple incarnation of the 'stairmaster fallacy': there are no links at the end of any essay that will lead you straight on to the next... you have to go 'upstairs' to the ToC and then back down.
It seems from the contents of this issue that articles neither need to be composed as hypermedia nor to discuss hypermedia from a Joycean perspective. Some of the articles, in fact, are reprints without even the virtue of 1st-pub originality. The editors might stress that electronic publications work much better with short paragraphs-- several of the essays fill the screen with solid blocks of unbroken text.
There's a topological oddity even in the "General Information" entry of the contents page: its contents (more a masthead than an intro) are repeated exactly, at the bottom of the contents page!
This sort of 'category error' is a weakness found everywhere in the pages of site designer, Rob Callahan. It seems to be a sort of link-drunkenness, for he rarely misses an opportunity to implement a link, no matter how many times it's been offered elsewhere in his pages. (One exception is the unlinked list of contributors.)
If your browser, like Netscape, is able to indicate links that you've already visited, this is not such a serious problem. But Lynx-users are being continually challenged to peek at files they can't be sure don't contain new material-- another version of the classic "gotcha". This can be minimized if files are organized neatly, without redundant linking.
Rob's pages overall are in much better shape than the journal, but he falls into several familiar traps:
#-links (to another anchor on the same page) should always be flagged as such, or they'll promote disorientation. HyperTerrorist recommends using "#" as the flag in some way, and "^#" if it leads 'backwards' in the file. Examples: (Mike#) (Skeffington#)
Unannotated hotlists
(if it's worth linking to, you should say why you liked it)
Here and
here.
Undescriptive anchor/link text (doesn't explain what it links to)
(obscure JJ-quotes for titles)
('Netscape')
('Free Mumia')
Content hidden in a hierarchy
(not just a timeline)
Too-long lists of items that need internal sorting
Here and
here and
here.
"False alarm" links (that warn you off a link you could follow)
(not just soundfiles)
Images missing ALT text (results in [IMAGE] or [INLINE] or [LINK])
[all over the place]
Missed opportunities for live links
(GNN, PC Mag)
Email address without live 'mailto'
Here
Time-references 'stale'
"most recent JJQ"
No indication of bottom of page (not important, but nice)
[all pages]
Other examples of Rob's characteristic category errors:
this page
offers direct
links to various resources that are repeated in
the 'onsite' list which
includes, in its own strange-loop, another link to
the 'offsite' list.
The timeline page includes not only a lot of GIF-gotchas and some undeclared #-links, but also a massive quantity of multiple links to the same documents. (Every reference to "Ulysses" need not be a link to the same file!)
Even more self-referentially, Rob's curiculum vitae has a link to itself at the top of the page. (This should be disabled.)
And, finally, along with a great number of other self-advertisements, Rob calls his site "The James Joyce Homepage". (Ahem.)
Special topics:
surfing-skills :
url-hacking :
open content :
semantics :
pagelength :
linktext :
startpages :
bookmarklets :
weblogging :
colors :
autobiographical pages :
thumbnail-graphics :
web-video :
timeline of hypertext
Anti-XML/W3C/etc:
structure-myth :
page-parsing :
firstcut-parser :
html-history :
semantic web
Design prototypes:
topical portal :
dense-content faq :
annotated lit :
random-access lit-summary :
poetry sampler :
gossipy history :
author-resources :
hyperlinked-timeline :
horizontal-timeslice :
web-dossier
Website-resource pages:
RobotWisdom.com :
Altavista.com :
1911encyclopedia.com :
Google.com :
IMDb.com :
Perseus.org :
Salon.com :
Yahoo.com
Older stuff:
design-lab :
design-checklist :
HyperTerrorist :
design-theory :
design cog-sci
Hosting provided by instinct.org. Content may be copied under Open Web Content License.