- It kind of works.
- It sucks your soul dry, which in turn limits its effectiveness.
DBSlayer lets you access your database via http and returns JSON. That's supposed to help with scaling. What I don't understand is this: won't accessing your db over http make things slow? Since you have to go over the internets? Compared with hitting your db machine that sits next to you in a rack?
Google widget: click to throw food to the fish.
Yesterday: "Louis Vitale, 75, a Franciscan priest, and Steve Kelly, 58, a Jesuit
priest, were each sentenced today to five months in federal prison for
attempting to deliver a letter opposing the teaching of torture at Fort
Huachuca in Arizona. Both priests were taken directly into jail from
the courtroom after sentencing."
Hey it's the 50th birthday of "satisficing".
I've started a consulting company to more explicitly focus on the stuff I've been doing for the past few years: help startups (or startup-like projects) develop consumer products, and help companies organize global websites ("global information architecture"). The company is called 290s consulting. Go check out the new company blog to find out why, category geeks might like it (and the .com was available, for one).
If you need help tuning or improving your mySQL databases or queries, percona seems a good choice. I haven't worked with them but I like their site and their blogs. 200$/hour for an efficient mySQL performance audit isn't bad. I'll report back if I ever use them.
Social networks and the wireframe as a boundary object.
"The objects that mediate the ties between people" is a powerful concept. On LinkedIn, the "recommendation" is an object that truly connects people, creates lock-in and ads long-term value (more than the network itself, the "friend" connection, for example).
From that post: "Think about the object as the reason why people affiliate with each
specific other and not just anyone. For instance, if the object is a
job, it will connect me to one set of people whereas a date will link
me to a radically different group. This is common sense but
unfortunately it's not included in the image of the network diagram
that most people imagine when they hear the term 'social network.'"
Exactly. Social networks with objects that connect people are stronger, longer lasting and provide more value.
Social science has more to say about these objects: they can connect different domains of expertise. When they do that, they're called boundary objects: objects that are used by different communities, and each community attaches different meaning to the object.
The wireframe (an IA deliverable) is a boundary object, and this power to connect different groups (desingers, coders, business people) through a shared object that has different meanings for each group is (I think) one of the reasons why the practice of information architecture has been such a success.
How exactly boundary objects and social networks that connect people through objects fit together I'm not sure. We'll figure it out :)
Now that's writing with balls: just tell the truth: "We've stopped issuing new Ethnio accounts until November while we make the product amazing instead of promising but buggy."
More on Dewey Decimal and its biases.
Reddit says: "The reason we didn't display more than 100 comments in the past and
still don't display more than 100 now is because of bandwidth and
rendering time (both ours and yours)."
I think that's a mistake. First: bandwidth of a zipped list of a few 100 comments is VERY small: a few K only. Smaller than 1 image often. Second, displaying a long page with comments is fast: here's an example (380+ comments). I am thinking about the same issue (you have to cap it somewhere, no point in displaying 1000 comments on 1 page?), but as long as there are no threads that hit 1000 comments, I think I won't bother. I don't think the reddit solution of only showing "good" comments is right. It breaks the conversation flow.
Lately I've been obsessed with the power of lists and leaderboards. I think smart lists are very, very powerful and this is something that can make or break a website. There's almost a book in there, but I'll keep that one for a rainy day :) Here's a good description by Lucas of the (ex) webJay algorythm, including a discussion on feedback loops (what's popular stays popular because it's on the most popular list).
The "Star Wars" project is alive and well: "The glossary of acronyms provided by the Pentagon to students of
missile defense is a list of abbreviations, like SBX or MIRACL (Mid
Infrared Advanced Chemical Laser). The glossary is typed on letter-size
paper with single-spaced entries and common type size. It is 327 pages
long."
Wrong strategy LinkedIn!
LinkedIn's "platform" will reportedly have to "approve" all apps that get on it, in order to make sure it stays nice and business-y.
Wrong strategy. Having to be approved will keep away developers. I understand the need to keep things clean and business-y, but there are much better ways to do that. Putting in constraints *before* the app even gets in is stupid: it's shooting your own platform in its foot. They should look at the ecosystem as a social system: bad stuff will get in. Instead of heavy guarding at the gates, they should smartly police inside, plus have clever encouragement for the types of apps they want in their system (featured app lists, make sure leaderboards for apps are weighed the right way, etc...). Please LinkedIn, we need Facebook competitors. Don't shoot yourself in the foot taking the "easy" solution to the app-moderation problem.
(Enough mixed methaphors for you?)
I wish Google analytics would do 1 more thing: let me log "events": short text messages with dates, so I can say: "added this feature on this date", or "moved to different host" or "was featured in the NYT", and so on. And then let me display these events against the stats, to easily see where they may have affected traffic. Easy to implement, not too hard to adjust the UI, come on guys! Pulease.
wullfmorgenthaler makes me laugh daily: (and has an RSS feed):
Unfuddle (via Simon) seems like a perfect fit for startups. Project management, bug tracking and source control, and it's got a free version too.
Is there any way that, using jQuery and the Taconite plugin, you can return HTML that's not valid XML? I tried using cdata, but it doesn't seem to work.
The amazing power of sequel on rails.
Achieve your childhood dreams
Great presentation on achieving your childhood dreams. Mine were:- Write a book. Check.
- My face on a stamp. Check.
- Man of the year in Time magazine. Check.
- Nobel prize for peace. (This one might be a bit harder).
- Get interviewed in Humo. (Not there yet, but one day.)
jquery question
I have a very strange jQuery problem: stuff works fine on my local install, but doesn't work on the live install. For example, sign up at poorbuthappy.com, then go to http://poorbuthappy.com/test/ and click the "add friend" button. It should be ajaxy, but it's not (well, kinda, you'll see. The button just greys out, and it should change to a "remove friend" button). The PHP script is reached, but the jQuery "success" function isn't fired. And I have other local Jquery bits that work locally but not live, while others do. It's weird.. any ideas?
When Mozy starts backing up (2 gigs of automatic backup for free!), at some point it says "reticulating splines". What does that mean?
Scancafe sounds pretty good for the photographers among us: you send them your old slides and negatives, and they send back a DVD with scanned images, high quality. For 25$ you can get about 100 slides scanned.
Digg's dead.
When you actually look at the stuff on Digg's homepage (today), it's boooring. Dead, filled with "7 ways to X" and similar SEO spam. Digg's out, as far as I can tell. The Yahoo homepage has a few similar problems, btw. (Specifically, check the URL's. The spammyness jumps out.)
Strč prst skrz krk is a Czech sentence. Sounds like something from a Superman comic (who was that little guy that spoke like that?)
I moved my book website from iabook.com to petervandijck.com/iabook/ to avoid problems with domain name squatters.
My sites on Mediatemple suddenly got REAL slow a few days ago. Luckily they fixed the problem with more hardware, memory and some smart tricks, and the good side of this is that now those sites are really fast. Let's hope they keep it that way.
It has become more and more clear to me that the biggest problem enterprises face in their IT selection is the software vendors, who try to sell monolithic "solution" that almost never delivers as promised. There is an inherent conflict of interest. Regardless of the advantages of a closely controlled environment, it's almost bound to go wrong, because there is so much incentive to make it go wrong. So what's the solution? Loosely coupled services that work well and are adopted bottom up can take care of a large part of an enterprise's software needs. For the bits that do need to be closely controlled, I guess we'll need the SAP's of this world for this for a while longer.
(Damn these cryptic posts with no examples! Oh well, no time.)
Google is pointing search traffic to their own properties, and now Yahoo's following suit.
I wonder why nobody has written about Google poisoning their search results with video links that without fail point back to Youtube, a video company that they own. And now Yahoo is doing the same: adding "useful" results to search, that point back to their own properties. Have the search engines decided that objectivity is no longer required? I always thought getting into the content game (buying Youtube) was a mistake for Google. Don't get me wrong though, I think the evolving "mixed" search results are brilliant, I just think that the search engine that makes an effort to send traffic to properties they don't own will have a much better chance of winning.
(I also think Yahoo needs to totally redo the yahoo.com homepage. It drives a lot of traffic I'm sure, but it's really quite bad. I'll keep that analysis for another post.)
Blip.tv's learning section is great. Really well done, I hadn't seen it before. Makes me want to do some videoblogging again :)
And indeed, "Homeland", just the term, is worrying. And that's after they changed that logo. "Because we don't learn much about European history, the setting up of a
department of "homeland" security - remember who else was keen on the
word "homeland" - didn't raise the alarm bells it might have."
Admit it: parallels with fascism in today's America are strong. It may be hard to notice over a latte in Sanfran, discussing Facebook or something, but come in fresh from the outside, and it's all there. Not that the US is fascist or totalitarian, but edging in that general direction? Yep. And that's a big deal.
In Europe, the memory of WW2 is still strong. In the US, I don't get that feeling. "First they came for the Jews." Most Americans don't understand yet
that the destruction of the rule of law at Guantánamo set a dangerous
precedent for them, too."
What is the next podcast brand?
With Odeo sold to Sonic mountain, the Yahoo podcast ghost town being closed, Podshow financially unviable (I believe so anyway, they raised way too much money and executed poorly), there's no strong brand in podcasting - except for iTunes. And iTunes doesn't really care about podcasting. Which means there's an opportunity there. The podcast format is still brilliant, and there are millions of unserved users out there (it's still not easy to make or find podcasts).
I was thinking what shape a successful podcast service could take. Creating podcasts still sucks, so that could be part of it. Finding them as well. I'm not sure if you'd need to go in the hardware business - it's possible.
I'm quite pleased I got to present my global IA talk at Drupalcon. Cross-fertilization between different groups is always a good thing.
It's the Halocaust! (video)
Designing from the core: sketchcast about information architecture ;)
I saw this great presentation called Core and Paths: Designing from the inside and out, at euroia, by Are Halland, NetLife Research, and i'm playing with sketchcast, and it had the simplest diagram ever, so I'm trying to recapture that here.
Hey, here's my first sketchcast. This *is* cool!
Users rate some sites consistently slow and other sites consistently fast. For example, About.com is rated slow. But when looking at the data, the sites that were rated slow weren't all really slow. Amazon is rated as fast, and it isn't really that fast (fast enough, but not crazy fast). So why do users feel some sites are slow or fast? (About.com loaded faster than Amazon.com in the test, yet was perceived as very slow.)
This is exactly the kind of stuff that makes me like Jared Spool.
The 12 habits of slightly successful bloggers.
- We write non-original stuff that we just read somewhere else,
- and then we forget to link to the original post,
- while also not doing any additional research
- and clicking POST really fast
- without thinking things through, which might have given us something interesting to say,
- and without proofraeding
- nor adding any interesting links and/or pictures, because really, this whole web thing isn't going anywhere, can we just go back to print?
- What's more, when we do want some traffic, we copy the style of successful writers without copying the content,
- and we SEO optimize the hell out of things, because we think that perhaps, yes, we aren't very good writers or original thinkers and perhaps no, we don't have much to say after all, deep in our hearts we know this, but since there are quite a few SEO blogs out there we have started to think that getting traffic isn't, somehow, about providing value,
- and what we also do is write really long posts sometimes,rambling on and on and on,
- making the same points again and again, until even the most bored, desparate, i-really-have-to-kill-this-time-because-i-despise-my-boss who-is-dating-that-cute-girl-in-accounting-instead-of-me and-who-does-she-really-think-she-is reader that may have accidentally stumbled upon our post has to just quit reading because it's just TOO much of the same damn crap that's all over the inkernets already and who the fuck do I have to kill to make it all stop?
(ps: and we can't count either.)
The Origin of Brands Blog: Apple's Golden Goose
The Origin of Brands Blog: Apple's Golden Goose: "Distractions are not what Apple needs. A focus is what they need. The way to build a monster brand is to attain global dominance. Which is what Nokia did in cellphones, Red Bull did in energy drinks and Google did in search."
Laura Ries is one of the most important branding thinkers in the world, and she nails it again on Apple: refocus on the iPod! The iPhone is, from a branding perspective, a distraction.
7 reasons I switched back to PHP after 2 years on Rails - O'Reilly Ruby
7 reasons I switched back to PHP after 2 years on Rails - O'Reilly Ruby: "I threw away 2 years of Rails code, and opened a new empty Subversion respository. Then in a mere TWO MONTHS, by myself, not even telling anyone I was doing this, using nothing but vi, and no frameworks, I rewrote CD Baby from scratch in PHP. Done! Launched! And it works amazingly well. It’s the most beautiful PHP I’ve ever written, all wonderfully MVC and DRY, and and I owe it all to Rails."
The EuroIA conference in Barcelona was wonderful, I'll do some follow-up posts soon.
blog.pmarca.com: The three kinds of platforms you meet on the Internet
blog.pmarca.com: The three kinds of platforms you meet on the Internet: "Like my old boss Jim Barksdale used to say, it's the difference between the chicken and the pig at a ham and egg breakfast. The chicken's involved but the pig is committed."
IA for beginners: take your time
For people who love words: http://wordie.org/ Check it out!
For a week of hardcore Drupal, head to Barcelona right now!