Views from Medina Road

I promise not to become preoccupied with this website Hugo debate, and I hadn't intended to post anything about it to SMOFs at all -- it's not really my place -- but I couldn't help responding to this exchange between David Dyer-Bennet and Leroy Berven over the issue I numbered second in the previous post: the recurring idea that one website Hugo category isn't enough, because big guns like Scifi.com, or Locus Online, which for whatever reason are perceived as having unfair advantage over more 'fannish' websites, would dominate the category.

>Quoth David Dyer-Bennet:
>
>> By the way, how do you propose to distinguish "fan" from "pro" sites?
>>
>> How would you categorize the following: www.tor.com, www.baen.com,
>> dragaera.info, www.neilgaiman.com, www.dendarii.com, isfdb.tamu.edu,
>> www.locusmag.com.
>>
>> (Off the top of my head, I'd categorize them as pro, pro, fan, pro,
>> fan, fan, semi-pro.
>
>So would I.
>
>> Locusmag would go to full pro if there had to be
>> only two categories. But I don't know what the rule is that sorts
>> them that way.)
>
>Given my druthers, my own preference would be that we not _have_ a need
>to develop one. OTOH, distinguishing "fan" from "pro" seems to be a
>process best conducted in the spirit of the U.S. Supreme Court
>justice's well-known definition of obscenity.
>
>Leroy Berven

I responded that it would require curious, counter-intuitive definitions of "semi-professional" and "professional" to nominate Locus Magazine as a "semi-pro" magazine --though it provides full-time employment for a publisher, an editor, and several full-time staff-- while designating Locus Online a "professional website" --though the income it brings in barely covers site expenses, and certainly doesn't provide full-time employment for anyone.

Though I do understand the desire for some distinction. Sponsorship, or affiliation, perhaps? SciFi Channel supports SciFi.com, and Locus Magazine supports Locus Online by virtues of name, information, and contacts, even if not any significant money. Best Affiliated Website and Best Independent Website, perhaps, analogous to Adapted and Original Screenplays?

As I've typed this post, David Dyer-Bennet has responded to the listserve with a detailed response that finds my suggestion unpersuasive. Oh well. Just a thought. I should stay out of this.

Since I'm *not* a SMOF, I had to be clued in yesterday by Cheryl Morgan about the debate brewing among SMOFS over the prospects for a revived website Hugo category. The debate hinges on several issues-- 1) whether there aren't already enough, or too many, Hugo categories; 2) whether a distinction can or should be made between funded, 'professional' sites and fan-run sites (where the latter are those of most concern to some debaters); and 3) whether it's possible to judge websites based on their appearance and content at the time of voting rather than what they were like during the year of eligibility.

I was queried about the second issue, since there is an ongoing, understandable impression that Locus Online must be a 'well-funded' site sponsored by Locus Magazine in the same way that SciFi.com is supported by the Sci Fi Channel. The truth is far different, as I responded. But I was more fascinated by the arguments over the third issue -- many posters suggesting that websites can't be judged, because websites are changing all the time. You can't see what the site was really like last year, because it might have been changed since then.

Well, yes. Websites aren't like print publications; and the ways they're not are what make them interesting and valuable, it seems to me. If this is a problem, it's a problem for the way awards categories are defined, not the websites. (I'm sure the SMOFs, being creative and forward-thinking like all skiffy folk, will work it out.) Anyway, the charge isn't usually true; most websites do let you see what the site was like in the past, because they archive their material. Here at Locus Online we even take 'snapshot' captures of our homepage once in a while -- note the 'Homepage capture' links at the top of the 2003 Archive page. Sites that remove old material, or implement complete site redesigns that affect archived material (via style sheets, say), are relatively few.

But this brings to mind the larger question of how to judge websites -- aside from how they're funded, or who's running them. It can't be purely content, or a plain text, static webpage of your favorite book or magazine would automatically be your favorite website.

I've always thought that there are three ways in which websites can -- and I'd suggest, should -- take advantage of the medium of the web.

1) Websites can update continuously. There's no reason to gather updates of a website into periodic 'issues'; that concept is a holdover from print media, useful for production scheduling and indexing perhaps, but otherwise unnecessary and anachronistic (analogous, I've always thought, to early films that were photographed versions of stage plays).

2) Websites can link to each other. Related material is just a click away -- and the selection of links is just as much an editorial decision as the selection of the site's unique content. There are still some traditional journalist sites that won't link out (for fear of losing readers) or who try to control links in (for advertising purposes, usually), but they too are anachronisms.

3) Websites can accumulate content. Not just new pieces of material, but growth of the entire site -- in this way, database and index sites (Imdb, the Locus Index to SF Awards) aren't like magazines, they're like reference books that don't require you to purchase new expanded editions every year.

There are certainly other criteria for evaluating websites -- content (of course), design and ease of use -- but the 3 principles listed above are those that characterize the web and the most popular sites on the web -- news sites, shopping sites, blogs, message boards. However beautiful their design or interesting their content, sites that don't take advantage of these capabilities might as well be... magazines. Or books. (Not that there's anything wrong with that.)

I've been preoccupied for the past 10 days or so reading short fiction published in 2003 in answer to an invitation to submit nominations for this year's Theodore Sturgeon Memorial Award.

I read lots of short fiction and reviewed it monthly in Locus Magazine from 1988 through 2001. Partly due to fatigue and partly due to personal circumstances, I virtually stopped after that. I realized last week that it had been fully 2 years since I'd read a story in a magazine. It's been oddly surrealistic to watch the parade of awards nominations and winners pass, this last year, without having read any of them. Circumstances having changed and improved, I hope to do some catching up this year, and to get back to keeping up at least somewhat, at least to the extent of reading the stories that others recommend, or that are nominated for the various awards. More on the experience of reading cold, vs what others have recommended, later.

Meanwhile, on the subject of voting, I was especially struck by this Slate article about psychology experiments and the recent history of Democratic primaries. It's for this reason that, despite some recent posts about the way Locus Poll votes accumulate, I would never hint about the results of those votes.

239 ballots received so far, including some paper ballots that a Locus staffperson is transcribing into the online form, so that they all become electronic entries into my Locus Poll tallying database...

I'm not particularly aggressive about promoting Locus Online or obsessive about its usage statistics (since unlike Locus Magazine, Locus Online is remotely far from supporting anyone, however modestly), so when a couple days ago I stumbled upon a web search and ranking utility called Alexa, I had the suspicion that I'd belatedly discovered something that most netizens are already aware of. After all, there was Alexa already built into my IE toolbar (under Tools/Show Related Links). I just hadn't clicked it before.

Doing so, I found that it shows the traffic rank of whatever site currently in the browser. So of course.. well, here's a ranking of sites I frequently visit, or was curious about, without further comment. The only caveat is that Alexa ranks by domain name, not specific URL, so I can't discover, for example, what the rank of Bruce Sterling's blog is by itself (it's part of wired.com), nor can I parse Locus Online from the Locus Index to SF or the Index to SF Awards.

Yahoo, MSN, Google, Ebay, Microsoft: 1-5
CNN	24
BBC	28
NY Times	66
Blogspot	406
LA Times	761
Salon	942
Wired	981
The Onion	1872
Scifi.com	3293
Andrew Sullivan	8,020
Fictionwise	11,068
Arts & Letters Daily	12,316
Popdex	20,635
Boing Boing	21,346
PW	34,299
Locus Online	69,358
Neil Gaiman	82,844
Danny Yee	89,421
SF Site	103,937
SFWA	113,973
Electrolite/Making Light	134,582
Baen WebScription	162,999
SF Crowsnest	207,960
ISFDB	215,996
Asimov's	470,818
Strange Horizons	485,973
Analog	632,765
The Alien Online	772,354
The Infinite Matrix	785,366
Tangent Online	1,300,900

195 ballots received thus far. There are people who have a tendency to detect patterns where none exist, but I try to resist such thinking, and so I'm sure it's mere coincidence that 15 or so ballots received in just the past couple days have voted for little else than the very same work in one of the novel categories. Others might be conspiracy-minded, but not me.

Two novels that I read recently both describe incidents of road rage. Being a southern California near-native, I'm alert to such issues. Each novel has a distinct attitude toward driving style.

In one [chapter 9], the protagonist recalls

living in San Francisco and hating every inch of the city, from the alleged pizza to the fucking! drivers!—in New York, the theory went, drivers used their horns by way of shouting "Ole!" as in, "Ole! You changed lanes!" "Ole! You cut me off!" "Ole! You're driving on the sidewalk!" while in San Francisco, a honking horn meant, "I wish you dead. Have a nice day. Dude."

while in the other [p113],

In general, drivers on the East Coast were less generous than Californians, Frank found. On the West Coast they played tit-for-tat, or even firm-but-fair, because it moved things along faster. Maybe this only meant Californians had lived through that many more freeway traffic jams. People had learned the game from birth, sitting in their baby-seats, and so in California cars in two merging lanes would alternate like the halves of a zipper, at considerable speed, everyone trusting everyone else to know the game and play it right...

and later from the same book [p155]

North on the freeway, crowded but not impossibly so, people zipping along like starlings, following the flocking rules keep as far apart from the rest as possible and change speeds as little as possible. The best drivers in the world.

The identity of the books is left as an exercise for the reader.

I updated the Locus Index to Science Fiction Awards today, mostly to correct a problem that an alert visitor to the site emailed me about a couple weeks ago. Apparently when I folded in, last month, results of the Locus Online 1999 poll for all-time best novella, novelette, and short story, I'd used a preliminary set of results rather than the final set as posted. Or in importing records from one database to another, they'd gotten scrambled somehow. Or something. Anyway they weren't right, so I wanted to get them fixed. I also added the results for all-time anthology and collection, which I hadn't gotten last month.

While I was at it, I folded in the recent nominations for the Nebulas, the Clarke, the PKD, and the BSFAs, though not 3 or 4 other sets of nominations that have already been announced this year. I'll get to them.

Since the pages of the Awards Index site are so inter-linked, it would be risky to update just a portion. So I always re-generate the entire site, some 2500+ html pages. Takes 1/2 hour to an hour for the database to generate all the pages (depending on which computer I'm on), and a similar amount of time to upload them all to the server (depending on what connection I have.)

I realize that some of the global tallies are now no longer strictly correct; they compile recent nominations for awards whose winners have not yet been announced, while the tally breakdown assumes nominations are either winners or losers. If I continue to do interim updates of the site, I'll have to fix the tallies so they only include awards whose results have been announced.

I replaced the initial offsite link to SFWA's Nebula Ballot with an onsite news page listing the nominees (with a couple spelling corrections), along with statistics about past nominations and wins. If obsessive attention to statistics is good enough for Oscar prognosticators, it's good enough for us Nebula and Hugo watchers.

It's remarkable that 5 of last year's 7 winners (counting people, not categories) are nominated again this year. There are 16 first-time nominees this year, though that counts 9 from the script category--the others are Moon, Baker, Doctorow, Vukcevich, Brockmeier, Van Pelt, and Gloss.

None of the novelette nominees has won before.

Of nominees who've never won before, Jack McDevitt and Karen Joy Fowler lead in nominations, with 8 and 7 prior nominations respectively.

That 5 of the 6 best novel nominees on this year's final Nebula ballot are women! Is this unusual? Well, let's see. Last year it was 2 out of 6. (Begin here and follow 'Nebula thread' links at the top to previous years.) In '02, 3 out of 7. Then 3 out of 6, 2 out of 6, 3 out of 6, 4 out of 7, 3 out of 6, 1 out of 6 in '96, then 2 out of 7, 1 out of 5, 4 out of 6 in '93, 3 out of 6 in '92, again in '91. That's far enough for the moment...

140* ballots received. I tabulated the first 80 last Friday to see if any categories have obvious front-runners. A couple do.

One more voting observation. Many (I'm tempted to say most, but haven't actually tabulated numbers) ballots from recognizable professionals 1) tend to be among the majority of ballots, described in the previous post, which register votes for only a couple items, and 2) usually vote for works by themselves, or works they have some professional association with.

Or for works by friends or associates... which is where I was headed with the "Clades" title last time. There are subcommunities within the larger SF/F/H community--starting with the SF, the F, and the H communities, of course, but including much more specific groups of writers and readers who share interest in a particular type of fiction and a counterpart disinterest in any other kind. Some of these subgroups are well-known, and even nickname themselves...

This is one reason there are so many SF/F/H awards, with such specific criteria distinguishing them. The awards that attempt to encompass the entire field--the Hugos, Nebulas, the Locus Poll--inevitably reflect compromises in their results, though perhaps not as seriously as in awards that limit voters to one choice per category. Example, heard this morning on the car radio: Robert Christgau explaining Coldplay's record of the year Grammy win. It was the only non-rap record among the finalists, so all the voters who hate rap voting for that. Speaking of the Grammys, my impression is that the music industry, and music fans, are even more balkanized than readers and literary types; most people have very specific musical tastes defined as much by the many styles they don't like, as by the few they do. The Grammys aren't so much one award with 100 categories, as a collection of awards for several dozen styles of music all grouped together under a single name. Now there's an idea for a way to overhaul the messy SF awards scene...

Or here's an idea. Perhaps we could simplify the Locus Poll. We could invite one group of voters to submit ballots--say, voters whose last names begin with A, or voters from a particular state. Tabulate their votes, and then eliminate any candidates with less than some significant percentage of the total. They would bow to inevitability and resign. After a few rounds of this, the winners will have been virtually determined, and the remaining voters can save themselves the trouble of actually voting. (If they do vote, they would merely be affirming the preselected winner, as in elections in some third-world nations.) Think of all the time and energy this would save.

*correction, 6p.m.; earlier I wrote "100 or so"

66 ballots received so far.

Another observation. Which isn't new news; you can deduce this from the statistics from past years (number of ballots, number of votes per category). Most ballots register votes for only one or two novels, and that's all (aside from survey answers). A minority, say one out of ten, or even fewer, submit votes in all five spaces for sf novel, all five for fantasy novel, all five for novella, etc. More like one out of twenty. Nearer the end I'll calculate actual statistics.

Another Locus Online special year-in-review piece, by Cynthia Ward, is in edit and will be posted shortly.

The online ballot for the Locus Poll and Survey is up and running, and already over a dozen ballots have been received. Voting is open for nearly three full months, so there's no reason to vote right away, especially since readers are encouraged to catch up on books and stories they might have missed--using, of course, Locus' handy Recommended Reading List, and the reviewer essays that appear in the February issue--before voting.

Whether due to the posting of the ballot, or more likely of the Recommended Reading List, traffic to Locus Online surged on Tuesday, to over twice the typical daily high-- over 18,000 unique visitors.

Votes in the poll generally trickle in at a steady rate from February, when the the ballot goes online, through the April or May deadline, with only a slight surge at the end as the deadline approaches. Having tabulated the votes the past 2 or 3 years (by importing the emailed form submissions into a database), I've noticed some interesting trends in how people vote, and how the votes accumulate. Without of course being too specific, I may mention some of these in the coming weeks as the ballots come in.

Here's one general observation to start. Though there are usually one or two close categories every year, in most categories in most years the eventual winner is apparent after only 40 or 50 ballots have been received. (The grand total being 5 or 600 or more.) Seeing this happen makes me much less skeptical about political polls, which are often criticized for using such small samples. (It's not the sample size, it's the selection.) That doesn't mean you shouldn't bother to vote unless you can vote early, of course. You won't know which categories are close or not, and your vote might well determine a close category, even at the very end.

Update 7.30 p.m. PST: about 40 ballots received, so far.

The online ballot for the Locus Poll and Survey is ready and was posted briefly last night before I realized that none of the test submissions I sent were appearing in my inbox. So I removed the homepage link and went investigating at my hosting service (that's CI Host), and quickly found this notice:

We regret to inform our Customers that the dp_tellafriend, cgiemail, insecure formmail, totmail, and alienform scripts have been disabled on all servers due to continual abuse.

Naturally, they had not notified their Customers of this, beyond posting the notice on their own website.

The key to part of the solution is 'insecure'. So the link to the poll form has to go through the secure server (i.e., https:// [etc]). Did that, submitted more tests, still didn't work. Further investigation uncovered revisions to the formmail cgi script, requiring the recipient email address in the form to explicitly match the email, or at least the format of the email, specified in the cgi script. All of these changes are spam-preventative measures, supposedly. Anyway, I made further changes, and am awaiting confirmation from Locus HQ that the latest test submissions got to where they're supposed to go. When everything looks OK, the ballot form will be back online.

I'm 90% over the flu, though it apparently has a long trailing edge.