“imagery”

Can’t believe what a coward I’ve been. I guess it wasn’t until I read recently about how Jon Lech Johansen humilliated Hollywood by releasing a program to easily break DVD’s much vaunted DRM (when he was 15 years old), or about “the college droputs that run >youTVpc.com for millions of people on just two low-end desktop computers that I realized how big a wuss I am.

Take Imagery. In a way, I never wanted it to be very popular because I knew I was doing something maybe-legal by scraping Google. I wanted to scrape Flickr and Yahoo! but feared they would be even less likely to see my scraping in a good light. I wanted to create a picture cache to make Imagery much faster and reliable (and to lessen the leech on independent websites) but fretted about bandwidth costs and about whether people would get angry about my caching.

Or take the Spanish dictionary of the Spanish Language Academy. It’s a tremendously useful, gratis dictionary and yet it has a butt-ugly, nonhumane interface. I’ve been bitching about it for years. And I’ve been thinking about giving it a new interface for that long, but, again, what if they don’t like my scraping? (In all probability they won’t. They’re the quintessential staid, monolithic organization.)

A lot of what ifs. But most importantly, so what? So sue me. Well, actually, most important is that I think these scrapings are an overwhelmingly good thing. I see them as a lot of fun—for me, for the scrapees (who’ll get to see how it’s done, ehem), and for the people who will enjoy just how much they never imagined missing. I see them as criticism by example.

So here’s Domburi and here’s NotReality. The idea of Domburi has changed somewhat from its inception (oh, the shame, so many goals not accomplished), it shall now be a collection of search superpowers instead of limiting itself to imagesearching. I believe, with an arrogance that I can’t believe but that I’ve missed, that I have a thing or two to teach Google in its own turf, not just in forgotten backwaters like imagesearching (where Imagery still kicks Google’s ass easily). So I’m starting really tiny now with only a simple redesign of Google search results (there isn’t even imagesearching yet). I’ll be fleshing it out in the coming days, daily, with the daylog here below.

Not Reality, otoh, is still what I’ve always wanted it to be: my webfront for my interface experiments. It’s really simple now but I’ll be improving it daily too.

So please visit the websites, leave your feedback in this post’s comments, and come back often to check out the daylogs—there are loads of interesting things in the pipeline! Thanks for reading.

12-14/May/07

Endless fiddling on Domburi. Collapse of the incrementalism. Obsession with pipe dreams.

15/May/07

Not Reality


Domburi

1. Ajaxed requests. Loading icon. Everything happens on the same page. Title changes. Very lightweight script behind the scenes: the now deprecated but still unsurpassed moo.ajax.
2. New one-column format with even results shaded. Results turn yellow on mouse over, which is silly interactivity, but surprisingly pleasant. Entire result is a link. No numbers anymore showing order.
3. Displayed URL is now a link that shows results only within the website. This idea from SearchMash, of which I just found out yesterday, and which is a website run by Google where it tries out new interface ideas without the Google brand skewing perception. Very intriguing.
4. Displayed URL is now cased smartly. So instead of greysanatomyinsider.com you get GreysAnatomyInsider.com. It extracts the case from the result’s title, uses some general heuristics (like upcasing after a ), and if all fails, it simply capitalizes.
5. Results displayed in-page. This had been the idea ever since I decided to expand Domburi from image searching. I wanted to make the whole searching experience feel faster and more like what Ben Schneiderman calls direct manipulation. It has been much harder than I thought and it was this point where I spent most of my fiddling (I also played for hours with the two-columns layout…). Here’s what I ended up with.
   
   1. Results seem to open in the entire page, with only the result entry on top. They really open in a full-screen iframe but the effect is surprising. To return to the results you can simply scroll out of the iframe, click the result entry, or…
   2. Pixelside. This is a strange but crucial feature that even if invisible and initially nonintuitive I find very, very promising: it’s simply a 1px-wide, 100%-height leftmost line. Right now you can click on it when on a full-screen result and return to the resultlist (and from there you can click again on it to return to your full-screen result). It is incredibly fast (in Fitt’s law^WP terms, it has “infinite size”), handy, and habituating—and I imagine lots of cool ways to enhance the feature. The only problem is that crappy IE doesn’t allow leftmost pixels! So I’ll have to make up for it with JS. I cringe with only the thought.
   3. The title (but not the entire result, which is JS triggered) is a normal link so if you want to open results in a separate tab just middle-click or CTRL-click it.
6. Strange cool script used. More details to follow. If you’re interested, check out $.
7. Works in Firefox, IE, and Opera. Not tested yet on Safari. Some weird bugs on Safari but it broadly works.

16/May/07—20/Jun/07

10-day trip to the US (haven’t told you about that!) with days way too happily busy. Too many books afterwards (63!) to do anything but read for several wonderful, obsessed days until all the stress and overcrowding of the house finally bring me down. Languishment in captivity.

21/Jun/07

Not Reality

Domburi

# Fixed logo to point correctly at home. Thanks volve!
# CSS fiddling for greater clarity, minor improvements, and cross-broswser compatibility.
# Added Google Analytics tracking.
# Height adjustment menu. A simple (though surprisingly troublesome) addition that makes Domburi much more useful, you can now adjust the height of the embedded windows and so can view  and compare two or three results at the same time! Imagine this with dragging and width-adjustment…

22 and 23/jun/07

Bad time management. Sorry. :)

24/jun/07

Domburi

Not Reality # New Book Section! With great quotes and cool photocovers.

Y’know, I remember reading about the syndrome in Brooks’s book with a smug confidence that it would never happen to me. It did. Imagery was by many accounts a pretty cool thing, but then I tried to outdo myself with its successor, Domburi, and, many, many ineffectual months later, I must admit that I’ve only weird sketches and weirder code to show for my time.

Which doesn’t mean that I’ve given up. It means that we need a new strategy. The all-or-nothing, hail-mary, next-big-thing, under-wraps-until-perfect approach was doomed since the beginning. (I really should have known better.) So the new strategy is to get it all out. As rough and soon as possible.

I’m calling it ”Improv’d Daily!” and it is akin to beta-hood^WP—in that it indicates that the website is still under developement—but it carries the all important mantra of radical incrementalism: every single day there will be at least one new, stand-alone, non-trivial improvement for the website. It won’t be earth shattering every day but it shall always be interesting.

I’m starting the meme with this very blog, which is supposed to be my online self and yet still lags far, far behind of what I want from it. (Domburi will be up in a couple of hours. Domburi up.) This very post will be updated daily with each day’s changes starting now and I have several new goodies to kickstart the kaizen:

8/May/07

# Related Posts section added (when viewing an individual post). Posts are related the more tags they have in common and the more rare those tags are.
# List of comments (accessible from the right sidebar, at the bottom of the Recent Comments header)
# New URLs: http://elzr.com/articles/YEAR/MONTH/DAY/TITLE becomes http://elzr.com/posts/TITLE, which is shorter and sweeter. You don’t need to remember a post’s date now and, what’s more, if there’s no post found with that TITLE, Google comes automagically to the rescue.
# Left sidebar redesign: new headshot, shorter description, just email (putting my phone # up there was always a bad idea, that phone-call confirmed it), new format for the archives.
# Collapsed “for:” tags in a post’s tag list. Much clearer. Tags are also now ordered alphabetically.
# Lots of tiny improvements all over. Like the orange bar atop a single post—neat, huh?—or icons for search (a magnifying glass in the searchbox) and for favorites (a star in favorite articles).


9/May/07

# Crappy day: a minor, bureaucratic improvement to the website became a nightmare. Blog crashing on and off. Domburi will have to wait until tomorrow.


10/May/07

# Blog back!
# Section Cache!: the recent list (favorites, posts, comments), the tags list, and the archive are now cached, making the website much, much faster.
# List of all posts (accessible from the left sidebar, below the Archives header)

11/May/07

# Save to Del.icio.us, Reddit, Digg, and Stumble Upon when viewing an individual post.
# Tag Cloud!
# js-less Improv’d Daily! Ok, this may not sound like much but it’s important and cool. I use ALA’s CSS Sprites technique.

12-14/May/07

Obsessed Domburi fiddling. Sorry.

15/May/07

# Fixed broken Tag Cloud links (Thanks Aaron!)

16/May/07—20/Jun/07

Big, humongous gap—or vacations—or depression bout. Or all of them together. See chronicle on Domburi’s Improv’d Daily.

21/Jun/07

1. Old URLs redirect to URLs to keep with the migration announced May 8. http://elzr.com/articles/YEAR/MONTH/DAY/TITLE now really becomes http://elzr.com/posts/TITLE.>
2. Sidebar Redesign: new picture, new welcome copy bared down to its barest Basic English^WP essentials, new webapps added to webapp section, new, much better descriptions for most items in the sidebar.
3. Daily Improves section in the sidebar for you to keep handy track of my progress—or lack thereof.
4. Minor CSS fiddling—like a new, bigger size for small caps type (it could be hard to read at some resolutions and some platforms).
5. New 404 page, that is, a new page to aid you when you type in an address that can’t be found. Try it now with http://elzr.com/this-address-is-wrong/. Thanks Aaron!
6. 
   
   

   New title for homepage. Since the delta thing is already obscure conceit enough, I decided to convert seconds into more humane time units. 8,321,231s delta is now 96 day delta.

   Delta, btw, means something like the divergence (the difference) that has come to pass between two different times, one of which is usually the present—so when I say in this blog’s homepage title that there’s a 96 day delta I mean that I haven’t updated it in 96 days, i.e., me and my digital self have had 96 days to go our own separate ways. This wonderful sense of the word comes from Charles Stross’s Accelerando.
   
   
7. Unified search into a simple URL, http://elzr.com/search/QUERY, which currently carries a personalized Google search of elzr.com but will eventually change to Domburi. This new unified interface allowed me to finally create a YubNub command for the blog: try elzr (see its man page) at every input box that speaks YubNub.

That weird phone call I got weeks ago was from the prim (but ambitious) lil’ supermarket near my house (they got my number from my blog, go figure). Out of the blue they demanded, not rudely but not friendly either, my “cooperation” in taking down pictures of them I had uploaded to Flickr (for I wanted to write a review of how innovative and important the store really is—“The income level of a country is determined, above all, by the productivity of its largest industries. High productivity in the unglamorous “old-economy” sectors—retailing, wholesaling, construction—is most important, since more people work in them.”). Anyway, it sure looked like a big boatload of crap to me then. They weren’t giving me even hints of good reasons and still they threatened me—me, their most ardent former enthusiast—that they didn’t want to pursue the matter in a different way (wtf?). I instinctively groped for the freedom-speech martyr role, willing to fight the crusade against dimwitted, Pleistocene shopkeepers to its bitter end.

And so it would have likely been. But then father and Dragonball intervened^ELZR. “If you do something that you later find upsets a friend, what you do is stop,” was father’s simple but crushing argument. Dragonball’s was more subtle in its nonverbalness but you could word it into this feel-good motto: “enemies are future friends waiting to be made.” I’ve never kept enemies and so it simply kills me to have one. I can’t. Because even if they never actively hurt me, I’ve always been aware that there will come a time when their help would come in handy—and I need all the help I can muster. In the case of this shoppe, I saw them immediately as customers. If this harebrained scheme I hatch of creating an ad-based online interface to Guadalajara is ever going to take off, I will need the help and patronage of every local business I can find.

It took me weeks to visit them (see previous fear post) but when I did, yesterday, it couldn’t have gone better. I went there and defused the whole thing by admitting error from the very beginning and promising to take down the pictures as soon as I came back home (which I’ve done). What followed was two persons trying to outapologize each other. My caller revealed himself a friendly, good-natured man. Most importantly, I finally got to understand what got them so upset. To begin with, being somewhat new to the retail business they’re paranoid about security after lots of bad experiences and it totally unnerved them when this random guy was able to sneak behind guards (some of them undercover) and take pictures nonchalantly. The crux of the matter, though, was that it turns out my Flickr page was the first Google hit for the store (that happens a lot whenever I talk about something from Mexico, Google gives me a totally disproportionate pride of place—Imagery’s aftermath, I suppose) and that, combined with the anxiety of having problems with their webdevelopers (who haven’t been able to upload anything—not even a lousy banner—in six months), got them all worked up—how can it be that some random stranger is the one that tells the world what we are?

Now I offered myself up for the job and I may be the one building their web presence, which I’m sure would be a fascinating job. Amazing isn’t it?

I doubt someone would find this too useful but I smiled today when I found about the guess YubNub command. You feed it text, it gulps the language it’s in. A great way to showcase YubNub’s open-ended fun, courtesy of Xerox research. It would have been a godsend when I was dealing with Imagery’s multilingual rush (Oh, how GMail angered me then! Smart enough to correctly spellcheck anything I gave her, yet coyly keeping the language name to herself!). Hope I need it again soon.

For all of you that aren’t on the YubNub wagon yet, you can play with it here—but it won’t be even half as much fun ;).

And since we already seem to be on a language landslide, some months ago I found out playing with Google Translate that when you translate a website from Chinese to English (which is currently beta), you can hover on a sentence to get the original Chinese fragment in a quick popup. Mighty cool. All the more impressive a feature coming from a website. (Now let’s only hope they plan to add it to the other language pairs too…)

Final language tidbit: translate “Hello, how are you?” to Spanish with Google. Your immediate response is “¿Hola, cómo eres?,” sucking the life out of even the hardiest machine-translation enthusiast.

The weird thing about fetishy pics^ELZR like the one below is they’re so beyond-words good if you were actually on the look for them you wouldn’t find any. This one, for instance, I found yesterday (here) while looking for pictures of sakura^WP, cherry blossoms, on Imagery.

In what must be one of its most bizarre moves to date, Google just released a collaborative-tagging game (!): Google Images Labeler. It frankly seems against the company’s algorithmic DNA and I almost dismissed it at first, but perhaps it’ll work… for a while: it’s actually interesting to play but the interest fades quickly. (Via John Battelle)

In a surprising study, Jakob Nielsen found out most (85%) of the learning gleaned from usability tests came from your first 5 user tests; 3 users tests will still give you a whopping 67% of the learning there is. He then goes on to advocate small and iterative testing over a single massive one (say, three tests with 3 users each and improvements in between instead of one lone test with 9):

Combine that with Joshua Schachter’s, founder of Del.icio.us, idea of “Starbucks usability tests” (“offer someone coffee to sit down and play with your product.”) and you’ll know why I’ll be this Friday on Minerva’s Starbucks harassing three every pretty girl I can find with Domburi (TeenVogue recently had a feature on young fashionistas being excellent for usability tests, can’t find the link though).

This (anonymous) feedback on Imagery just came on Saturday.

Some sort of auto-citation of images is a fantastic idea (as anyone who uses EverNote or Google Notebook will know firsthand) and my gratitude goes to whoever sent it to me, I’d never have thought of it myself. And yet, for a while I almost decided to willfully not implement it:

I strongly disagree with the way citations are usually handled within “education-land”: little more than curtsies one must mindlessly perform to pay respect to others’ property (and it is against such moralistic establishment that I am one of those kids who believes he owns the internet). Citation styles are taught and required simply as one more formal hoop for students to jump.

But citations can be much more than that! They allow readers to recover and rewalk the path the writer followed, and in that they perform an invaluable service to readers, but they can also be immensely profitable for writers too, starting with forcing them to walk paths in the first place (one is so loathe to do the slightest of researches when in the thrall (or duty) of writing, so very prone to simply rearrange one’s prejudices and call it even). Citations make for more rigorous reading and writing—that’s why we should encourage them (not simply because they make, arguably, good fences).

So yes, I thought I saw some of that ownership-based, rote teaching of citations (copyright-instruction) in that email—in a scared flash of exaggeration I glimpsed a DRM image-search engine—and my recoil reaction was so surprisingly strong I thought of deliberately not implementing any sort of auto-quotation. Lawrence Lessig has talked already on the power technology’s architecture has to regulate conduct and the weight of such responsibility was suddenly overwhelming.

Careful thought has shown me the error of my ways. My overreaction to such friendly (and helpful) feedback was not called for. An auto-citation feature in Domburi would be very helpful indeed and will be implemented. But it’ll be tinged with my prejudices and that means it will be open-ended.

Interesting: hundreds of people are just pouring to Imagery from, of all places, Slovakia^WP. This Slovakian website apperas to be the cause. How weird (I did nothing particular for this to happen) and how cool (it’s the 21st language on which there’s written record of Imagery talk!).

Imagery’s multilingual feedback^ELZR has been the best, most rewarding part of it all. I was feeling down with Domburi the other day (and with how hard it is to get the interface just right), but then a new review came in and things are bright and beaming again: