association-list

For those of you unfamiliar with the competition, the John W. Campbell Memorial Award is a juried best novel prize given out each year. There are rankings, 1st, 2nd and 3rd place prizes are awarded. The jury changes a little each year, it seems and thus things are a little mixed, in terms of prediction. They definitely seem to consider all of the major novels each year, and it looks like they try to throw in a few books that wouldn’t normally be considered as part of the SF mainstream (i.e. weren’t published by one of the standard SF imprints.

While I don’t always agree with their selections, there is a hint of awarding books based on merit rather than politics or popularity, which I can only applaud. I’m not typically one for media spectacle, and sometimes the Hugos and the Nebulas are irritating in that they seem to be overwhelmed by those factors. So, a few words about each of the finalists, followed by some predictions. But first, a retraction: mentioned in the last post that I’d talk more about Ian McDonald’s River of Gods in this post, but looking again, he doesn’t seem to be here, I’m guessing that it was published too late in the year and will likely make it onto the ballot next year (I also don’t know whether or not the selection committee is going off of US or UK release dates in the case that a novel had both, that also might factor in). In any case, he’s not here, and it’s a book that needs talking about. So, another post on McDonald, some day soon.

Transcendent, by Stephen Baxter

I haven’t read this one. It isn’t really my policy to read Baxter anymore. It’s all interesting big ideas stuff, but his writing never really grabbed me and the level of shared universe navel-gazing got too extreme at some point and I stopped reading.

The Meq, by Steve Cash

People seem to like this book, as they also seem to like mentioning that Cash was once part of the The Ozark Mountain Daredevils. The synopses and reviews that I’ve read lead me to believe that it isn’t something that I’d usually go for, but I’ll likely pick it up now that the original is in paperback, on the strength of positive buzz, and because I’m a weak, weak man when it comes to buying books.

Child Of Earth, By David Gerrold

I hadn’t heard of this one before today, to be honest (and many readers will be wondering why I bothered commenting on all of these if I haven’t read any of them. I’ve read most of the second half of the list, I promise). It looks like a YA SF novel, which is really very strange coming from Gerrold, who I know best from his unfinished War Against the Chtorr series, which are grim and violent and sexual to a degree which would make one thing that Gerrold is an unlikely children’s author. That said, Scott Westerfeld is writing YA now, and all of his books before Midnighters are grim and violent and sexual, so what do I know. Evolution’s Darling, one of his earliest, is a wonderfully strange novel full of post-human fucking and interclade love affairs, from what I remember. Neat stuff, if often uncomfortable. I saw him speak not too long ago and he admitted that the reasons for switching to YA are at least partially financial, which, in my mind, is a sad state of affairs. Not that I don’t like his YA books, but they’re forced by the tenor of the times to skirt too widely around too many issues for them to be entirely engaging. But I digress, wildly.

Mind’s Eye, By Paul McAuley

I like Paul McAuley. I haven’t read this, because it isn’t out here and I haven’t seen it at the store (they carry imports, but can’t really get all of them. I’ll have to ask). I have no idea if it’s any good, but I would assume so, based on his history.

Seeker, By Jack McDevitt

I’ve only read one book, Chindi, which I think that this is a follow up to. A lot of people seem to like McDevitt, but I don’t really see it. Despite the increasingly furious pace of socio-technical evolution, the people in these books seem already in the past, other than that they’ve got neato spaceships and know how to go faster than light. It makes any speculation that they books may put forth seem strangely stunted. Also the writing is lackluster and the characters seem stock and the whole thing could have been written 20 years ago. I’m still talking about Chindi, mind you. I haven’t read this one, and some trivial research indicates that it’s entirely unrelated and seemingly more up my alley than that book. So, we’ll see. When it comes out in paperback. Maybe.

Learning The World, By Ken MacLeod

OK. Now we’re back on familiar ground. Ken MacLeod is one of my favorite authors from the past couple of years, and this is one of his best books. Rarely, in the newly burgeoning field of singularity related SF, do you see anything positive. We’re destroyed or irrevocably transformed or fractionated or herded into easily manageable groups. It’s always a disaster. MacLeod thinks differently. In his future, we’ve survived, Singularity is a chronic but tractable infection and mankind has spread to the stars in a big, big way. You can see human space for light years as the stars go green with life. Megastructures of truly astonishing scale are mooted. Travel tubes between solar systems? Why not? Immortality? Old hat. In my opinion this is one of the best visions I’ve seen of a humanity truly triumphant over dumb matter, not huddling in tiny colonies on hostile world, fighting grim, useless battles with incomprehensibly alien stand-ins for earthly foes. Not here. Space is too big and too rich to fight over. Planets are pleasant places to evolve and grow up, but a mature civilization has too much on the move to be bothered with such limitations. Throw in an interesting first contact story cum detective tale and you’ve got one hell of a novel.

The Summer Isles, By Ian R. MacLeod

Our second MacLeod, Ian, is a gifted writer. He lacks some of the speculative brio of many other prominent SF writers, but he has a true gift for evoking atmosphere, and that’s usually enough. He’s much like Gene Wolfe in that respect, which is high praise. I’m actually not sure whether I’ve read this or not. I’ve read the story of the same name in one of his other collections, but it looks like this has been extended out to the length of a short novel, so it’s quite likely that I’ve only read part of it. In any case, it’s very interesting, a story of an England that lost WWI and went fascist immediately (rather than waiting for Thatcher), and there are many parallels with early Nazi Germany, as is standard. The story left me wanting to know more, so I’ll have to pick up the novel if indeed it’s been extended. Although it looks like it’s going to be an expensive book to come by.

Counting Heads, By David Marusek

Counting Heads, at first, looks like it’s going to be as packed with wonderful and bizarre speculation as Stross’s Accelerando, but there are long gaps in the pace of innovation and there are pacing issues where Marusek seems to just settle down to actually tell the story that’s going on here, and doesn’t quite achieve the balance that he’s looking for. I liked it, don’t get me wrong, but it seems a story unfinished and one that never quite pays off fully. David Marusek is a fucking crackerjack short story writer, and I’m looking forward to the followup, as hopefully he’ll get more comfortable with the form and start cranking out some incredible novels.

Mindscan, By Robert J. Sawyer

Didn’t read this one. Never been a huge fan of Sawyer’s, and the book didn’t really seem all that interesting to me, as it seems to me that a lot of the issues therein have been richly addressed in the past.

Accelerando, By Charles Stross

Hard to say anything about this one that hasn’t already been said, but that never stopped me before. A string of novellas fix-uped into a novel with, as far as I could tell, very little alteration. Explosively inventive to the point that it’s hard to follow occasionally, as people change and mutate and are manipulated by a weakly godlike feline AI. Even if all of the stories aren’t great and it doesn’t really hang together all that well as a cohesive novel, this should be required reading if you really want to know what’s possible and what’s going on in SF today.

The World Before, By Karen Traviss

I’ve not gotten to this one yet. I thought that City of Pearl, the first book in the trilogy of which The World Before is the third novel, was pretty good, and very interesting as a first novel. Traviss isn’t pushing a ton of boundaries, but she’s a good writer and draws interesting and sympathetic characters.

Spin, by Robert Charles Wilson

I avoided Robert Charles Wilson for years because I confused him with Robert Anton Wilson, author of the Illuminatus! books. Oh lordy was I wrong. Once I realized that, I read everything of his that I could get my hands on. Wilson is a fine, fine writer of prose, and never fails to tackle Big Ideas. Sometimes he fails, but that’s a good thing. His forte seems to be tracking in great and telling detail the reactions of humans to events quite beyond their ken. A truly enjoyable read and one that only lets you down a little when the ending doesn’t quite live up to the promise of the start, but you don’t really care because the writing is too wonderful to regret.

Evan’s Hopeful Predicitons:

1. Learning the World

2. Spin

3. Accelerando

Evan’s Cynical Predictions:

1. Mindscan

2. Spin

3. Seeker

Also, I note that I’ve been linked to by Jeremy, which means that there might be more than just two or three of you out there now (gotta love RSS readers). Just as a note, I close comments because I don’t have time to deal the the comment spam, nor do I have the time to learn the ins and outs of wordpress’s spam reductions mechanisms. However, I always enjoy getting feedback and hearing what other people have to say about the issues that I’m addressing. There’s an email contact link at the bottom of the page (obfuscated a tad for spammers, but simple enough), and I’d love to hear from you.

Living Next Door to the God of Love, by Justina Robson

Ignoring the misstep that was Natural History (it wasn’t bad, it just wasn’t as good as this one or Silver Screen), everything that Robson does seems to get both weirded and better. I am avidly awaiting her next novel. Really, go out and read this one, and pick up Silver Screen while you’re there.

Macrolife, by George Zebrowski

Pyr confuses me. On one hand, they publish things like Silver Screen and River of Gods, the latter of which is easily one of the best SF novels to have come out in the last couple of years (I’ll talk about this one more in the context of the Campbell jury prize nominees). On the other hand, the spent part of their budget bringing back Macrolife, which, while it isn’t by any means a horrible book, and is in fact filled with many wild and wonderful ideas, I’d rather have just read a pamphlet of the ideas and skipped the bullshit cozy catastrophe story about an American immigrant dynasty (we’re still pretty white-bread), who accidentally destroy the world and then make good anyway. The second section is equally painful story wise, but you’re already past the new ideas until the third part, so what you get is some hand wringing about the First World’s abandonment of the Third World and then a justification that it would hinder growth too much to stop and help everyone (anyone?) else out, and to stop growing is to die, etc. Admittedly, this came out when I was a year old, so perhaps I’m missing some historical context, or things have changed in the interim, but top this off with an astonishingly grating intro, and you’re left with a book that has some great ideas that are accompanied with a more interesting setting and much better writing in Iain M. Banks’ Culture novels.

The Overnight, by Ramsay Campbell

This is the first horror novel that I’ve read in a while (meaning around ten years). Incredibly creepy and atmospheric, but tends to rely on the same tricks over and over again. Ultimately, it ends up being incredibly grim. Worth it for the creep out at the time, though.

Visionary in Residence, by Bruce Sterling

For my money, the best of cyberpunk to come out of the original movement (stories aside) in the ’80s were Schizmatrix, Vacuum Flowers and Count Zero (flame away). Ever since then, Bruce Sterling has been jerking me around. Time to re-re-re-read A Good Old-Fashioned Future or Holy Fire. Chairman Omniveritas has not delivered today.

The Voyage of the Sable Keech, by Neal Asher

I am a shameless Asher fanboy. Read The Skinner read it all! NOWNOWNOW!!one1

The Bonehunters, by Stephen Erikson

I’m not usually a fan of the huge, never-ending fantasy sagas. So it is with vast embarrassment that I admit that I bought this at great expense as soon as Borderlands imported me a copy from the UK and I enjoyed it very much, thank you.

The Ghost Brigades, by John Scalzi

I’m waiting for The Lost Colony before I write about this whole sequence. Fun reading. I think that, while the antecedents to this and Old Man’s War are fairly plain, people do Scalzi a disservice my comparing him with Heinlein. Mouthpiece characters and wish fulfilment sex are pretty much nowhere to be seen. I have the feeling that Scalzi, is one of those writers who is pretty good but makes a massive, mid-career jump upwards in quality (good->awesome), must like Vernor Vinge with A Fire Upon the Deep (you should read that too, post upcoming, eventually). I don’t know why I get this feeling, but there, I’ve said it.

Ring of Swords, by Eleanor Arnason

I enjoyed this, but I cannot claim to be totally convinced by her argument and or analysis of sex relations. She seems to avoid all of the interesting stuff that’s referred to in the novel. Although it all makes sense on the level of story, it often fails to excite.

The Etched City, by K. J. Bishop

I’m assuming that Bishop is part of the whole interstitial fiction crowd and this may or may not be true. Regardless, her stuff reads like their stuff, but while most of it leaves me cold, I quite liked this one. It put me in mind of One Hundred Years of Solitude. It’s stuffed full of gorgeous description and telling details and calamitous and incomprehensible actions. What you were looking for if you were thrilled, then disappointed by Veniss Underground

Darkland, by Liz Williams

I am also a Liz Williams fanboy. She can do no wrong. You should buy ten copies of everything she’s ever written and give all but one copy of each to people that you know.

Godplayers, by Damien Broderick

Meh. Kind of a less morally ambiguous update of Players at the Game of People. Feels… I dunno, smug. Broderick isn’t nearly as clever or as original as he seems to think he is. I’d give it a miss.

Dusk, by Tim Lebbon

My first encounter with Tim Lebbon. Quite good, but I’m thinking that I’ll have to wait for the continuation before I can make any calls about where this is really going. John Clute says it better and more elliptically than I can in this brief space. I’m anxiously waiting for the sequel, but for right now, I’m going to dig through his apparently fantastic back catalogue (because I don’t already have enough books on the To Read shelf).

Reading Currently:

Double Vision, by Tricia Sullivan

Recently Read:

Living Next Door to the God of Love, by Justina Robson

Macrolife, by George Zebrowski

The Overnight, by Ramsay Campbell

London Revenant, by Conrad Williams

Visionary in Residence, by Bruce Sterling

The Voyage of the Sable Keech, by Neal Asher

The Bonehunters, by Stephen Erikson

The Ghost Brigades, by John Scalzi

Ring of Swords, by Eleanor Arnason

The Etched City, by K. J. Bishop

Darkland, by Liz Williams

Godplayers, by Damien Broderick

Dusk, by Tim Lebbon

On the To-Read Shelf:

A * means partially read.

The Prestige, by Christopher Priest

The Separation, by Christopher Priest

The Female Man,* by Joanna Russ

The Game Players of Titan, by Philip K. Dick

Radio Free Albemuth, by Philip K. Dick

Worldwired, by Elizabeth Bear

The Facts of Life, by Graham Joyce

The Demon Princes books 3-5, by Jack Vance

Igniting the Reaches,* by David Drake

Midnighters book 1, by Scott Westerfeld

The Dark Beyond the Stars, by Frank M Robinson

Wild Things,* by Charles Coleman Finlay

The Prodigal Troll, by Charles Coleman Finlay

City Come a Walkin’, by John Shirley

A Fire in the Sun, by Alec Effinger

Effendi, by John Courtenay Grimwood

Predator’s Gold by Philip Reeve

Castles Made of Sand, by Gwyneth Jones

Life,* by Gwyneth Jones

The City of Saints and Madmen, by Jeff Vandermeer

The Lost District,* by Joel Lane

King of Morning, Queen of Day,* by Ian McDonald

Crossing the Line, by Karen Traviss

A Quantum Murder,* by Peter F. Hamilton

Context by John Meany

Spares, by Michael Marshall Smith

Bad Voltage, by Jonathan Littell

The Ghost Sister, by Liz Williams

Mothership, by John Brosnan

The City, Not Long After, by Pat Murphy

Also about six months of F&SF and Asimov’s.

I really do. Sadly, most of his work, save for the quite good but not nearly the best Bones of the Earth, are out of print. This is a crime. So there will be little linking in this post, because there’s really nothing to link to. Go to your used bookstore to seek this stuff out, or check online.

His novelistic career began with In the Drift which is an ambiguous little book about fanaticism in a post-apocalyptic landscape, set close to home (he lives in Philadelphia). Three Mile Island has melted down and poisoned the landscape. Weird Things Happen. It’s his first novel. I read it a long time ago and it didn’t leave much of a mark.

He hits his stride with Vacuum Flowers, a novel written at more or less the height of cyberpunk, but more full of more interesting ideas than anything other than perhaps Schimatrix, by Bruce Sterling (you should read that, too). Swanwick explores here what it might mean for humanity when the brain is a known quantity. Earth, the planet, has been lost to Earth, the hive mind, know as the Comprise. Everyone else lives in space, because speed of light jitter keeps Earth from branching out too far. The story revolves around a Mysterious Quality held by the by the implanted personality of the viewpoint character. To figure it out, everyone and their dog and their runaway planet will Stop At Nothing to get a hold of her and cut up her brain to figure it out. She runs and she hides, and therein lies the real show. You, the reader, get an awesome guided tour of the survivor state left by the birth of god, where people change personalities like clothes and make deals with the devil (Earth again) to make their deadlines. You meet a man with perhaps the most interesting case of self-inflicted multiple personalities committed to the page. Tyler Durden really has nothing on Wyeth. The characters here are weird, as Swanwick is in full bore Othering mode, forcing the reader to try and understand the experiences of people whose lives are barely comprehensible if you stop to think about them, which, given the short length and rapid pace of the novel, you’re rarely inclined to do. For all that, they’re warm and human and you care about them, although you might not entirely understand their concerns. After all, these are people who alter and replace bits of their minds like the characters in a Warren Ellis comic treat their limbs and sexual organs.

From there, we take a mighty leap forwards time and upwards in vision and style to Stations of the Tide. The book won the Nebula for best novel in 1991, apropos of nothing, other than that it’s out of print now, which, to repeat myself, is a crime against all that’s good and pleasant in the world. The narrative details the journey of an nameless bureaucrat sent by the Department of Technology Transfer to recover some potentially stolen technology on the surface of an embargoed world called Miranda, where the natives may or may not be sentient, and they may or may not be extinct, and most all of the animals have an aquatic for and a form for land, because the unexplained Jubilee Tides overwhelm most of the world every so often. The book notably contains many nods to previous works of science fiction, which honestly are its weakest point. The real meat of the story concerns the bureaucrat’s internal journey towards understanding his and his Department’s role the hinted-at interstellar polity which provides the frame for the story. Earth is mentioned only once, but significantly. The majority of his story is couched in the chaotic fin de siecle ambiance with something of the rotting glamour of New Orleans as it’s often seen in literature combined with the wild west enthusiasm of a dying town where law and order have generally already packed up left. I am completely failing to capture here how thoroughly and convincingly Swanwick establishes and maintains this atmosphere. Every detail, from the most mundane to the almost painfully surreal, helps to further embed the reader into the weave of the story. Again, the trip is the thing, here. The solution to the central mystery is almost secondary by the time you reach it, for you are enthralled first and foremost with the bureaucrat’s internal crisis and debate over a change of direction. Stations of the Tide is definitely one of the better SF books writting in the last 20 years, and would be on any of the lists I keep threatening to compile about what you should read.

The Iron Dragon’s Daughter looks like it’s getting a hardcover re-release sometime this year. This is a good thing, but honestly it’s more than a little weird, as this isn’t a book that’s easy to market. There’s no cover image yet, and I imagine that it’ll be at least as bad as the cover for Neal Asher’s The Skinner. As always, I’m totally willing to be proved wrong, but the current trend seems to point to covers that get worse and worse and worse and worse. Expect a separate post about that soon. Anyway. I can’t imagine how they’re going to market this, except as a small release so that all of us who have ratty paperback copies can buy something that might last more than another reading or two. It’s awfully weird and wonderful, emphasis on the weird, I suppose, for the general reader. There’s a seamless melding here of celtic inspired fantasy, contemporary fiction and science fiction. You never really get a clear idea of just how seriously the book is taking itself. It’s been quite a while since I’ve read it, because my copy has been passed from hand to hand to hand and it isn’t quite clear where it’s settled. I’ll re-read it soon, and perhaps update. It wasn’t as much to my taste as Stations of the Tide, but it’s huge and entertaining and filled to bursting with wonderful ideas and rich characters.

To be honest, I’m doing Swanwick a disservice by focusing on his novels, as excellent as they are. Some, perhaps most, of his best work is at shorter lenghts, and there are a couple of North Atlantic Books short story collections of his work, two of which are still available. Tales of Old Earth and Gravity’s Angels, if you live somewhere without a decent local bookstore who can do special orders for you. To be honest, I don’t own either of them, but I should and will and I feel that I can recommend them to you without reservation, as Swanwick’s short stories are generally excellent and it’s pretty magnificent to be able to pick up most of his short stories for less than 25 dollars. And remember, kids, when you’re short of cash, your local library will often surprise you with the breadth of its offerings.

Also worth checking out is his website, which is full of neat stuff.

For all of the longer posts on this site, I use emacs as a text editor (refill-mode is your friend). However, if you’re using Wordpress, which has some sort of wonky rich text editor on by default, the hard coded breaks tend to cause some problems. My suggestion is to turn off the rich text editor (which I don’t feel is all that helpful, you might think differently), which at the very bottom of the page when you click on the Users tab of your admin page, and download, install and activate the php-markdown plugin. It’s a world of improvement once you get used to the markdown syntax, which shouldn’t take long at all.

There are two M. John Harrison novels that you should read, more or less immediately, if you haven’t. The Centauri Device and Light. It’s an investment though, since Centauri Device is only available in a Gollancz (UK) SF Masterworks edition, so if you’re interested, read his story Tourism here, or in your copy of years best SF 22, which of course you own. If you like that, then the investment is worth it. That said, some people have hated the story and then loved the books, so your mileage may vary. It doesn’t really matter, though, because if you don’t love M. John Harrison, you’re a bad person. Fundamentally and without question.

I’d read _Light_ before, a couple of years ago, on the strength of a recommendation by someone, though I can’t recall who, although it was probably Jeremy Lassen, book pimp and co-owner of [Nightshade Books](http://nightshadebooks.com). Cheated a little, and bought the UK edition cheap out of Canada, since it wasn’t coming out here for ages. I thought that it was great then, but the qualities of the book didn’t really stick with me that first time. I kept recommending the book to people who would never read it unless I loaned it to them, which, I think, is how I lost my first copy. It’s out there somewhere, unless I got it from the library. That time period is a bit hazy with boredom and frustration. In any case, I bought it again the other day without hesitation. Harrison both is and isn’t one of my favorite writers, but mostly is. He writes about losers, mostly, and not the cute, cuddly kind that you can feel well about when they make good at the end, such as you might find in vintage Gibson and Sterling. Most of the people in his novels are sad figures with massive flaws as visible and obvious as suppurating sores, who generally manage to fuck up everyone around them in the long run, both their friends and their enemies, and in Harrison’s work, these categories are terribly fungible. So, losers, kicked around by forces that they can neither affect nor ultimately comprehend and being pushed towards some end that we cannot see or anticipate, with the reader barely stuck in place by a mixture of charismatic characterization and the sick fascination that rivets your attention to an accident in progress. It sounds a lot like the overarching themes of the cyberpunks, really and would be fairly flat at this point were not Harrison one of the better writers working today.

It’s interesting to compare the two books side by side. They were written in more or less the same milieu about twenty nine years apart. Harrison has written little else in this particular setting, but I think that he must think about it, now and then, because the universe that we wander into in The Centauri Device is rough, still a little just off the rack, and the world in Light is an old but sturdy thing, many-patched and full of details and utterly comfortable to wear. Device is told through a smudged lens, its narrator utterly reliable but whose deep sympathies his characters and for the story that he’s telling come through very clearly. Light doesn’t avail itself of unreliable techniques either, and it’s narrative voice couldn’t be more neutral, but it almost seems as if the backing behind the story has slipped, the ontology underneath itself become unreliable. You might think that I’m speaking of Dickian ontological riffs, but it isn’t really that. Reality in Light is just as banal as in the phenomenal universe, it’s just less… concrete. By looking, you find, even if what you’re looking for wasn’t there to be discovered when you started looking. It seems to me to be quantum theory taken as both narrative ethos and as a force deeply tied to the observer, whomever there is to observe. Every method of going faster than light is possible, if you try it. The universe spawns anxious, grandmotherly ghosts from our unease and gawky aliens from our frustration at our own awkwardnesses. The three interlocking stories tease the reader, worry at their own coherence constantly. The universe is tired and outsiderish, more than willing to play these nasty little jokes endlessly.

The Centauri Device is full of poetic graces and the intentionally harsh enjambments of sense words with exquisitely mistuned modifiers that the cyberpunks would later call ‘crammed prose’. There’s an excitement to it, a sense that there are new things to be done here, on this particular overtrod path. It’s a young man’s book and a book mired deep in the cold war, one primary tenet being that there’s a third nation of the sick and the tired who are weary of the war which is their environment entire. This doesn’t, I think, decrease its relevance, for all that the Cold War has been over for all of my adult life. There’s always a massive conflict with which you are peripherally involved that and you yearn for it to be over, no matter which side prevails. Backed up by the strength of the writing and the a science fictional background which refuses to take itself seriously. And it’s funny. I mean, what more can you ask for?

Light is a different matter entirely. You’ll laugh, sure, but there’s a grimness here that a few laughs cannot overcome. The character in Device were more lovable losers than the ones here, with the possible exception of Chinese Ed. A sense of fatalism and inevitability overhangs and cannot be dislodged. You get the feeling that when a character is less that well rounded, that it is intentional. They’re flat because they have nothing more to offer the world than they offer their reader. They are simply what they are and it isn’t enough. The writing, which is always important to consider, when you’re speaking of Harrison, because his plots do not alway enliven and his characters, while well limned, make you want to be sick most of the time, the writing is more understated, without the shimmer and the bounce and it doesn’t force your attention here and there, but you never forget that it’s there. By ‘modern’ standard of mimetic writing, I suppose that it’s something of a failure, because occasionally you’ll find yourself pulled out of the story and forced to re-read a line or or a choice phrase just because it’s so good. But I’ve never held with the theory that writing should be a pane through which a scene is viewed. That particular pile of horseshit has led to more boring writing about imagined affairs than I care to recall, and I generally shy away from mainstream literary fiction. Writing should suit the story and Light manages this effortlessly and admirably.

Both books have warts, and their subject matter isn’t exactly what many people will pick up for pleasure reading. They’re quite weird in a lot of ways, and their gaze is obsessive. They’re worth reading, though, without any sort of doubt at all. I recommend them to everyone (although a lot of people are stopped dead by the first chapter of Light, but you should press on). They are not art that comforts, but they can be learned from.

The rich text edit box for Firefox/Linux fucking sucks. Anyone know how to turn it off in Wordpress?

The idea is to write something that can persistently store values for applications where a database is overkill.

I suppose, ultimately, that my complaint isn’t with databases as an idea, it’s more with SQL and with their ubiquity and philosophy. the Berkeley database system is kind of like what I’m thinking of, architecturally, but I would like to allow both network and memory access to the database as well as allowing more dynamic querying of that Berkeley allows, but keeping the speed of something like BDB as much as possible. Also, it would be nice to see just how simple and primitive I can make the system, so it can be applied to as many problem domains as possible. In the end, it would be nice to have written something so simple and general that it’s quite easy to write a simple layer to get this system to act just like a SQL, Object, or Berkeley database.

Abstract.

Essentially there are a couple of standard, built-in types, from which all other types can be constructed, up to and including objects. It should be easy, at the end, to write a very thin layer for things like persistence of C++ or Python objects, but that isn’t the main goal. The main goal is to have a super quick and low resource server where you can store/retrieve data. Something like object prevalence layers for simple types, but a tad more robust (i.e. with built in networking and smart to-disk streaming).

The whole idea comes from hating databases and thinking that SQL sucks, so I’m not going to write any support for the most ‘advanced’ features of those systems. It’s overkill and really a bit too slow for a lot of high performance usage, and it makes it difficult to detect errors at runtime without a lot of expensive string parsing. There will be a per-thread errno and a ps_perror() which will explain, clearly, what went wrong, so the program using the store will be able to react appropriately.

Details.

Types:

• byte is the basic storage type. Anything can be expressed as an array of bytes, except perhaps booleans, which are a special case that I’m not really interested in supporting. although this will be the only type that’s defined at base level, there will be several ‘typedef-ed’ types available at the user level and for programming convenience
• int: name of a class of types including s8, u6, s16, u16, s32, u32, s64, u64, etc.. stored in a machine independent order that be the same as network order so we don’t have to translate them for sending over the wire, so can just send them out without translating the order.
• strings: arrays of bytes meant for holding characters. There will be subtypes for Unicode strings as well. strings over a certain size (1k or so) will be compressed by default.
• blobs: basically strings with no implied structure. an ordered array of unsigned bytes (or words, if that’s faster). also compressed by default.
• dates: a special type of u64 with some assumed information about the content, which will be the number of seconds from the standard Unix epoch.
• handles: a special type of u64 which refers to another object known about by the same server. These handles are permanent and unique, and used for fast access to stored data.

Compression and other speed stuff:

• going to use libLZF, because, as the author says, it’s essentially free on modern CPUs. All items will be stored in marshalled, compressed formats so that once they’re located in memory or on disk they can be sent or copied immediately.
• when the server and the client are running on the same machine, the client library should transparently move to faster ipc methods than sending packets over the loopback interface. otherwise, the client side should just use sockets to talk to the server.
• chatter between the client and the server should be in a highly optimized binary format.
• not sure about threaded .vs event driven. will at least need some form of threading so that network/IPC speed doesn’t lead to blocking on query fetches.
• the entire store set will be mmaped by default. This really isn’t meant to take more than a couple of gigs of data, and the entire working set should be storable in memory. This is easy to get around, though. Most of the huge databases that I’ve heard of are storing giant binary blobs, which you can work around by just writing them out as files and then storing the file handle, and using already in place raid and backup mechanisms for safety, since I don’t think that storing this stuff in a database vs. storing it on a file system buys you much.

Access:

• Should be able to define new types from primitive types on the client side. not sure whether this should happen on the client or server side. could just happen on the client side and then get transferred over to the server into a special store.
• queries are a sticky issue. not really interested in providing joins or other advanced and generally less useful-that-you-think database features. The idea here is that CMS of some sort should be able to be implemented without all of the pain of setting up and tuning a database, although, ultimately it stems from my desire to write MMO server systems that don’t use databases. Not that I do that for a living, but, hey, a man can dream.
• a store call on the client side, if successful, should return an 8-byte handle (drawn from a pool of handles that are basically u64 integers). This should allow for space effective client side caching of fetched information.
• labels will also be allowed, which will allow a-list type access for searching for handles that you don’t know about. unlabeled items will not have a space in the store’s label hash, and cannot be searched for, except in relation to structures that hold the handle for those objects.
• editing structures to have new members and migrating all of the current data over should be easy, although it might be space expensive. There are a lot of issues here that would be best addressed after the process of writing the app has started, though, so it isn’t entirely useful to speculate on it now.
• transactions and other functionality can be simulated by requesting locks. I don’t want to build this in, really, as I think that it’s an incorrect assumption that everything will need to be accessed by multiple clients. Stores may or may not be atomic by default, I’ve yet to decide.
• relational store/fetch (relational in terms of structure) should be easy.
• freeing memory allocated by the client is always the caller’s responsibility.
• finally, named stores, which I feel that I should mention since I’m trying to be exhaustive, though it would seem intuitive.

Examples:

• The client creates (or perhaps has predefined) a structure called ‘character’. This structure has several members and goes in a store of the same name. Each of the members, e.g. hp, class, mana, history, etc., will be a handle to some value in another store. So you would then be able to do a fetch(char_store_handle, “label”, &character_struct); and the client library would, if it returns without error, fetch all of the information about that character and mutate the character_struct structure with the information that it fetches. It would then store the handle for fast store access when it needs to update or delete that structure.
• if the client wanted to sort all of the characters on the server by max_hp, say, it would then call a multi item fetch on the character store, grabbing the character’s name and the handle of the associated value within the max_hp store. from there, how you get to the data, sort it and then present it are up to you, on the theory that optimizing code is easier than optimizing database queries.