Monday, August 28, 2017

WorldCon 75 Takeaways

Eric and Greg inside the convention hall.
Two weeks ago, Eric and I attended WorldCon 75 in Helsinki, Finland. I'll discuss things of interest to readers of short fiction and fans of the Hugo Awards first, and I'll leave the personal stuff to the end.

The Hugo Awards

Winners

Our function at Rocket Stack Rank is to make it easier for people to find good short fiction and other information that could help them make informed nominations for the Hugos and other awards. We are not here to tell people how to vote. For that reason, we didn't offer suggestions for the final vote--not even a "this is how we're voting" post. For the same reason, we aren't going to evaluate the stories that actually won the awards. We offer our congratulations to all the winners.

Losers

Greg with Lady Business' Susan & Ira at Hugo Reception.
Rocket Stack Rank itself was a nominee for Best Fanzine, finishing 5 out of 6 (never thought I’d be grateful to Vox Day). That’s not really a surprise; we’re still very new (not quite two years old yet) and most fans still don’t know who we are.

We offer our congratulations to Best Fanzine winner Lady Business and to all the other finalists.

Voting Rules Changes

Greg and George discuss 3SV.
The Business Meeting took a very conservative approach to the voting rules this year, voting down EPH+ and 3SV, which had been approved last year. They also had a chance to suspend EPH itself for next year, but that was defeated as well. As a result, nominations in 2018 will operate just as they did this year. (To see what those terms mean, read Fix the Slating Problem Forever.)

None of these votes were even close. This comes as no great surprise; people are tired of the whole issue, and most people seemed to feel that EPH had done a good enough job this year. Also, the Puppies seem to have run out of steam too. Judging from their turnout for the final vote, there are only about 30 of the Rabid Puppies left, and the Sad Puppies have been inactive for over a year now. That suggests that there might not even be dinosaur porn on the ballot at San Jose.

Length of Novella

Nicholas Whyte and I cosponsored a measure to simplify the eligibility requirement for Best Novella. Lots of people are aware that the eligibility rules in Section 3.2.8 of the WSFS Constitution  currently say that a work that’s overly long (or short) for a category can still be eligible if it’s within 20% of the required length. Not many people know that there’s an extra rule that only affects novellas that limits this margin to 5,000 words. We proposed to remove the special rule and just use 20% for all categories.

In practical terms, that means that although the limit for novella is 40,000 words, works up to 45,000 have been eligible in the past, but now works up to 48,000 would be eligible. If this had been in effect  this year, “Penric’s Mission” would have been eligible for best novella.

The real point of the change isn’t to very slightly extend the range of novellas, although that’s nice (there are almost no novels shorter than 90,000 words anymore, so upping the range for novellas makes sense); the point is to simplify the rules. I don’t know how many times I had to explain to people why “Penric’s Mission” wasn’t eligible this year.

Anyway, this one passed by acclamation. If it’s ratified in San Jose next year, the 5,000-word cap will be removed, affecting works published in 2018. The 2018 Hugos (awarded to works published in 2017) will still be under the old rules.

Rocket Stack Rank

If Today is Tuesday, This Must Be WorldCon

Eric and Greg in the fjords of Western Norway.
Like a lot of other people, we used the trip as an excuse to explore the Nordic countries, so by the time WorldCon75 started on August 8, we’d already been in Europe for 17 days and visited 7 foreign countries (if you count changing planes in Iceland). Then, almost immediately after we got back to Seattle, we packed up and drove down to Oregon to see the total eclipse of the sun. So from July 22 to August 25, we were pretty busy.

Meeting People

Last year, at MidAmeriCon II, we gave away a lot of ribbons, but almost no one had ever heard of us. This year, almost half the people we spoke to at least had a vague idea who we were. A couple of authors told me that they knew who we were because whenever they searched on Google for the title of their stories, RSR’s review came up first, so our SEO strategy seems to be working.

Because I had a bad cold the entire time, we didn’t do the Stroll with the Stars events nor did we attend any parties other than a brief appearance at the Hugo Losers’ Party, opting to go home early and sleep. We did attend the File770 events, and enjoyed meeting people we’d only ever seen online.

New Ideas

Greg with "Standback"
We always learn things from the people we talk to at conventions and we usually get ideas for things to change about Rocket Stack Rank

Message Fiction

Probably the largest change is that we're going to try to make an effort in our mini-reviews to acknowledge the message the author was trying to deliver. We'll still treat a non-subtle message as a negative factor, but we'll at least give the story credit for conveying that message. If nothing else, it'll eliminate cases where people thought that somehow we didn't get the message.

Mary-Sue Characters

To quote Django Wexler, "A Mary Sue is a character for whom author favoritism is so strong that it breaks suspension of disbelief." The young person who's an absolute expert on everything, instantly becomes adept at all new tasks, and solves all challenges in a paragraph or two is a Mary Sue. Occasionally you meet a "Typhoid Mary Sue," which is a villain who enjoys the same degree of improbable success.

Despite the name, most Mary-Sue characters are male. 

Our policy had been that a story with a Mary-Sue protagonist automatically got one star, no matter what other merits it might have, on the grounds that this is an elementary writing error, but from this point forward, we're going to treat it as a suspension-of-disbelief problem, which has a limit of two stars. That makes it more consistent with the way we treat non-protagonist Mary Sues and characters that are almost Mary Sues.

Panels

This year I sat on four panels, two of which I moderated.

Artificial intelligence in Real Life and SF

(L to R) S.B. Divya, Greg Hullender, Anthony Eichenlaub
The AI panel went very smoothly, partly because we’d been swapping e-mails for weeks beforehand and realized we were all in broad agreement: SF stories do a terrible job with AI.

The room was packed; every seat was filled five minutes before start time, and the officials turned people away at the doors.

People see a lot of advances in AI, but they don't realize that all of the products that actually do useful work are able to function because they avoid the need for intelligence. Think of the search engine that doesn't try to understand the sentences in a document--it just does statistics on the words in it. This has solved a lot of useful problems, but (we argued) it hasn't made any progress at all towards a machine you could hold a conversation with, which you see in SF all the time.

Apparently AI is a very popular topic.
"Real AI" (aka "General AI") needs some sort of breakthrough. It's impossible to say that such a breakthrough could never happen, but it's fair to say that, at the moment, our understanding of intelligence is so limited that we couldn't even say what the nature of that breakthrough might be.

This is my list of pet peeves about AI stories, which I shared with the other panelists before the event. I think we addressed all of them one way or another:
  1. The “emotional AI.” Easily gets choked up and ends up unable to complete important tasks. Even if you had the technology, why would you build such a thing? In stories, the emotions always cause problems—there is never any benefit from them. A special case is the AI that’s paranoid about being turned off. I always wonder how it feels about accepting patches.
  2. The super-fast AI. This is where the AI does its thinking at the speed modern computers do integer arithmetic. Especially when we’re told the AI is physically spread out all over the Earth and yet it thinks a nanosecond is a long time.
  3. The emergent AI. Merely connecting millions of machines together causes intelligence to emerge. Probably popular with authors who think that if you mix all the chemicals in a chemistry set together it’ll result in alien life.
  4. The AI that “violates its own programming.” Makes me think of a children’s story about a train that leaves the tracks and runs through the streets of town, astonishing the citizens (“where is that train going?”) in order to meet its friends (a bus and a bulldozer) in the park for a picnic.
  5. The singularity. The irrational belief that if an AI designs the next generation of AIs, the progress will be exponential and not just sigmoidal. 
  6. The unique AI. It can’t be backed up, so if anything goes wrong, it’s lost forever. 
  7. The evolving AI. It changes quickly in response to new data in the field, even though these changes (amounting to entire new features) haven’t ever been tested.
  8. The perfect AI. Never makes a mistake. Based on how humans react to it, it never has made a mistake ever.
We were asked how SF stories should present AI. We suggested presenting it as a tool, not a companion. A story that revolved around the need to get enough data to train an AI would be unique. Or the risks of letting an AI learn in the field. (Most AIs are trained in the factory and learn little or nothing in the field.)

The exception would be stories set far in the future—500 to 1000 years from now. In that case, the author shouldn’t try to explain anything about how the AI technology worked—certainly not in terms of any technology that exists today.

I expected someone in the audience to really resist the idea that real work in AI today isn’t going to lead to HAL 9000 in the next few years. That the AI that can drive your car still isn’t going to talk to you about anything else. Perhaps the unanimity of the panel was daunting.

Short Fiction

After last year’s infamous meltdown, this year’s short-story panel was very tame. We discussed where short fiction gets published (print magazines, free online magazines, anthologies, and stand-alone novellas), the importance of short stories (beginner writers get rejected a lot, and short stories are a good way to get rejections quickly), whether the field is declining (subscription magazines are, but novellas are having a golden age), and which stories and authors we’d recommend (RSR makes lists of recommended stories).

The panelists were all authors or reviewers, but there were no editors, which somewhat limited the discussion of  how well the SFF magazines are doing.

The State of Machine Translation

This was the first panel I ever moderated for an SF convention. (I’ve moderated panels at scientific conferences before though.) As with the AI panel, we’d had weeks to exchange e-mails, and, again, we were all on the same page: Machine translation isn’t going to be good enough to translate stories until the general AI problem is solved. (And see the AI panel above for the likelihood of that.)

It's worth mentioning that translation was a major track at this conference, with sixteen different panels regarding different aspects of it. I attended most of them and confirmed that none of the professional translators found current translation tools useful for translating fiction. Technical documents are another story, of course because they often contain boilerplate text that has to be translated the same way every time, and for that, tools are helpful.

So if it's not useful to the professionals and it's not useful to monolingual fans who just want to read and enjoy SFF written in foreign languages, is it of any use at all? It turns out that the existing tools do have value for people who can already read at an intermediate level but who occasionally need a second opinion on a difficult paragraph. And, of course, when you get away from reading fiction, they're handy for reading menus too.

Power of reviewing when it comes to promoting or hiding diverse voices

(L to R) Emma Humphries, Elizabeth Hand, Cheryl Morgan,
Erin Roberts, Greg Hullender
I worried about this panel for several reasons:
  • This was the very last panel on the very last day, and I worried that no one would show up.
  • This is a very sensitive topic; it would be very easy to see things getting nasty.
  • If it didn’t get nasty, it could get dull. What if we ran out of things to talk about and the audience didn’t have questions?
  • I was the only male on the panel, but I was the moderator. Not great optics.
  • The panelists were unable to communicate at all before the panel, so I had no idea what anyone wanted to talk about.
  • I was sick as a dog that day and would have skipped the Con entirely if I hadn’t had to moderate.
  • The panel was supposed to explore ideas in Joanna Russ’s “How to Suppress Women Writers,” but I thought the book was out-of-date. (E.g. its #1 concern was reviewers who would deny that a woman really wrote a given work.)
  • Then I forgot and left my notes and my copy of the book in our hotel room.
But it was fine. The room was enormous, but we still almost filled it up. No one but me had even bothered to (re)read the book for this, and those who had ever read it weren’t interested in talking about it, so we just never mentioned it. Everyone was polite and respectful. The panelists supported each other almost as strongly as on the panels that had had weeks of online discussion.

And the Sudafed and Ibuprofen started to work about 15 minutes into the panel.

The only issue we really disagreed on was on the question of authors writing about groups they don’t belong to. That is, a man writing a story with a woman protagonist. Or a white person writing about a black person. No one actually said that it's wrong to ever do that, but one panelist had the position that if you do that, you have to get it right. I disagreed, arguing that it didn’t have to be right, just as long as it was entertaining to people in the group portrayed. I gave the example of straight women who write about gay men. We actually like those stories, even though they often make us laugh and shake our heads. Someone in the audience then asked what I thought about a story in which a lesbian is raped and it magically turns her straight. I responded that that would destroy suspension of disbelief (never mind not being entertaining), and thus get a low review.

The funniest moment was when we discussed how to measure diversity in fiction. The challenge is that although you can often guess the gender of authors from their names, it's much harder to do that for race. When I said, "I just can't see writing letters to authors saying, 'Hi! Are you black?'" Erin leaned over and said, "I get that all the time."

Hugo Nomination

And the Best Fanzine Award for 2017 goes to Lady Business!
It was a real honor to be nominated for an award. We really want to thank all of our loyal fans. Keep reading short fiction!

As for what effect RSR is having on the nominations, that’s still difficult to measure.

It’s clear that no one is using us as a slate, though; some things we recommended against had no trouble getting nominated while other works we (and other reviewers) recommended highly didn’t even make the top 15. On the other hand, the list of nominees is overall much higher quality (i.e. much more likely to have a recommendation from someone) than a similar list of stories selected at random. That could well be simply because we and the fans are both identifying the best stories.

In other words, by our standards, good things are getting nominated (and winning awards), but it’s hard to say to what extent (if any) we’re affecting that. That’s probably a good thing. We’re not trying to control the nominations; we just want to make it easier for people to find SFF stories they like.

Conclusion

Things went well, and we're happy we went. We'll write a post or two analyzing the final statistics once we get the time. 

10 comments (may contain spoilers):

  1. My biggest disappointment was not meeting you all in Helsinki, as I relied on your website to find works to read, let alone nominate. Again thank you for your hard word making it easier for me.

    ReplyDelete
  2. Thanks for the report. Good photos too. Plus my RSR ribbons arrived home - thank-you. The ribbon has gone on the end of my Sasquan name-tag, as I didn't make it to Finland in person.

    I myself use RSR as a reading and buying guide, so you are having an impact there. I then filter independently from there for award nominations.

    As for stuff RSR recommends against, some stories just "HIT you emotionally" so ends up on getting nominated regardless of other possible issues that gets noticed. I think I have only nominated 1 story that RSR rated a 2, and your comments about it were valid. Never nominated a story RSR rated a 1.

    ReplyDelete
    Replies
    1. It was a pleasure meeting your husband. Perhaps we'll get a chance to meet you in San Jose or Dublin.

      Delete
    2. We are thinking about attending San Jose and Dublin, but it will be Dublin, IF we do go.

      Thinking about it, I think I nominated at least 2 (maybe 3) stories that RSR rated a 2.

      Delete
    3. I hope you commented on them! :-)

      Delete
  3. Rather than 5th of 6, I'd say it was fifth from a field of hundreds - getting a nom was an excellent result.

    The novella wordcount change is an excellent little idea - hopefully it'll get the nod through next year. It'll be interesting to see how far the novella length gets pushed by publishers - there's a recent tor.com release Starfire: A Red Peace that I clock at 53k, although they are describing it as a short novel rather than a novella.

    ReplyDelete
    Replies
    1. I'd be happy to see the novella limit move from 40K to 50K (which is 60K when you add 20%), on the grounds that 90K is a short novel these days.

      Delete
  4. The Business Meeting also created a committee to review Hugo categories, so we may get some proposals to fine-tune some of them. I would love to see the wordcounts change a bit to increase novella's maximum and novel's minimum. That's not in the amendments referred to them, but they are charged with reviewing the whole set, so you never know.

    ReplyDelete