EagerEyes Shorts

Keep Your Org-Chart to Yourself!

Thu, 11 Mar 2010 22:05:30 -0500

I was looking for Kaiser Fung’s Numbers Rule Your World on Amazon, and saw that there was no Kindle version. So I figured I’d check with the publisher, McGraw-Hill, if there was perhaps some other kind of e-book (like O’Reilly has for most of its titles).

Googling for the name McGraw-Hill took me to their Corporate Website, which has a convenient little search box in the upper right. So I typed in the title, but only got bogus results. Nothing seemed to match the full title, and the pages were all generic corporate-style stuff.

Turns out there’s a little drop-down box on that search results/refinement page where you can select what you want to search. It defaults to McGraw-Hill Corporate, but it has options like Books and Educational Products. And lo and behold, of course it finds the book once I tell it to search for books.

But why does it search the corporate site first? Isn’t it more likely that people will want to search for your products, rather than your press releases or investor info? You have a lot more customers than investors, so why does your search default to boring corporate stuff?

It’s not that there is an additional step involved here that I can’t be bothered with, it’s the attitude. I don’t care about your org chart or your subsidiaries or whatever. I want information, fast. I want the website that comes up when I search for your name to make sense to 90% of the people. How can you put so much effort into a website and not think about who the visitors are and what they’re looking for?

This is similar to HP’s “license plate” URLs, and I’ve seen IBM do the same (though they seem to have stopped doing it). Don’t tell me your internal structure or your server names. Don’t tell me what you think is important; I don’t care. I care about the stuff that’s important to me. I want a simple URL, I want relevant information.

Is that really too much to ask?

These scans from the Ninth U.S. Census 1870 are truly...

Thu, 11 Mar 2010 10:09:41 -0500

These scans from the Ninth U.S. Census 1870 are truly fascinating. Check out the fiscal chart on page 35, the population pyramids on pages 40 and 41, the death charts on pages 45-46, and the disabilities on pages 49ff (including “insanity” and “idiocy”). There are also a few more treemap-like pages like the one above.

Incredible to think that all of these were done long before computers, by hand. The number crunching alone must have been a huge task, and then drawing these charts and transferring them onto plates for printing. Wow.

(via robertogreco)

fuckyeahinfo: datavis: Why does a salad cost mmore than a Big...

Wed, 10 Mar 2010 11:24:38 -0500

fuckyeahinfo:

datavis:

Why does a salad cost mmore than a Big Mac?

It’s a good cause, but the chart badly distorts the data. See my recent March Chart Madness criticism roundup for an explanation (it’s about 2/3 down the page).

Interesting idea by Sam Loman: the human body as a subway map....

Thu, 04 Mar 2010 08:25:01 -0500

Interesting idea by Sam Loman: the human body as a subway map. Blood vessels, nerves, digestive system, etc.

This is fascinating and disturbing at the same time: a dance...

Wed, 03 Mar 2010 12:09:06 -0500

This is fascinating and disturbing at the same time: a dance piece that uses reactive graphics triggered by the dancers. It’s called Chunky Move.

roomthily: giantrobotlasers: lookatthisfuckinggraph: Why 3D...

Tue, 02 Mar 2010 20:11:09 -0500

roomthily:

giantrobotlasers:

lookatthisfuckinggraph:

Why 3D pie charts are bad « fox @ fury

it is an accurate chart

Incredible!

marco: Embezzle a bit more of your building’s heat. Here’s a...

Sat, 05 Dec 2009 19:10:00 -0500

marco:

Embezzle a bit more of your building’s heat.

Here’s a useful trick to unfairly get more heat into your apartment than the others in your building: Use a box fan to convert your passive radiator — in effect, a giant heatsink — into an active heatsink-with-fan combo.

Of course, you can also just remove the shelf that’s sitting on top of the radiator and blocks 90% of the (purely heat-induced) airflow. Less noise, too.

Penny Arcade

Fri, 27 Nov 2009 13:49:58 -0500

Penny Arcade

There has been an interesting decrease in attempted comment spam...

Sun, 15 Nov 2009 10:51:44 -0500

There has been an interesting decrease in attempted comment spam on my website over the last two weeks. This may have to do with the recent attack on the Mega-D botnet, which was effectively destroyed by security researchers. The possibility of taking out botnets has been debated for a while, but so far only as a theoretical idea. The results of this attack speak for itself, and I hope that this will be done more in the future.

The OpenOffice Mouse. This is exactly what’s wrong with...

Fri, 06 Nov 2009 19:58:39 -0500

The OpenOffice Mouse. This is exactly what’s wrong with usability and Open Source.

Unicorn Sales Dashboard. Via @r1c1 on Twitter.

Fri, 06 Nov 2009 19:52:08 -0500

Unicorn Sales Dashboard. Via @r1c1 on Twitter.

Photo

Sat, 26 Sep 2009 14:36:03 -0400

Photo

Sat, 26 Sep 2009 14:25:38 -0400

Combating Spam with Mollom and Comment Closer

Mon, 31 Aug 2009 22:48:10 -0400

Comment spam has been a part of running my website almost from day one. Until about a month ago, I used reCAPTCHA, which is a very clever implementation of a CAPTCHA that at the same time helps digitize scanned books.

It started with a few spam comments that got through, and developed into waves. I would get bursts of about a dozen comments in quick succession once or twice a day. I had turned comment moderation on at that point, but it was still annoying. Then the waves intensified to the point where I got several hundred spam comments per day. Something needed to be done.

I had looked at Mollom before, but didn’t want to hand over control over my comments to a third party. At first, it also didn’t work too well, letting about half the spam through. But it improved over time, and it’s rock-solid now, to the point where I have turned off comment moderation again. Most of my users never see a CAPTCHA and don’t have to wait for me to moderate their comments.

There are other spam block modules for Drupal, but Mollom is much better integrated than the others. It does pay to have Dries Buytaert, the founder of Drupal, also behind Mollom. The integrated comment moderation queue and the clever report interface are simply excellent. The only thing the report charts is missing are false negatives.

What has also helped is another module named Comment Closer. It simply deactivates comments on nodes that are older than a given threshold. I don’t generally like closing comments on older nodes, because I don’t see my articles as being so tied to the current moment that they will lose relevance very quickly. But experience also shows that I get hardly any comments on nodes that are a week or two old, so I’m not losing much. And reducing the number of targets for the spam bots has cut down the spam attempts by about 2/3.

It’s still shocking to see the force behind these spam attacks, but I’m also glad that there are tools that are keeping up and making it virtually impossible for that crap to be posted.

What is the best app for designing logos, etc?

Tue, 18 Aug 2009 14:27:50 -0400

I’m working on a redesign of EagerEyes, and while I have some reasonable tools, things still take way too long. I know that designers like PhotoShop, but I don’t have that, and it’s out of my price range. PhotoShop Elements is too limited in the functions I’d need (I have an older version of that that I never use).

My main tool right now is Acorn, which has most of the features I need, but it also lacks a lot of convenience. There is no way to group layers, making it hard to keep an overview when trying out lots of things. It also requires a few additional steps for inner shadows (which I may or may not use in my redesign ;), etc.

The tool that would be perfect for my purposes in theory is DrawIt, which is a vector-oriented tool that can also do pixel-stuff like apply CoreImage filters. It takes some getting used to, but it’s really powerful and has some incredible workflow features. The main downside is that its text rendering is really ugly, and it doesn’t look like there’s a fix in sight.

I also have a license for VectorDesigner, but there are some issues with getting that to really do what I want it to. I need to play with that a bit more, though.

Is there some other tool that’s used by designers, runs on a Mac, and is affordable? I need mostly pixel-oriented stuff, but a hybrid like DrawIt would be great, too. Any suggestions?

jeannr: I made a flow chart, that we might better...

Tue, 18 Aug 2009 12:40:14 -0400

jeannr:

I made a flow chart, that we might better understand.

Once upon a time I was falling in love, and now I’m only falling apart.

Brilliant. (via @ryanmcgrath on Twitter)

Imogen Heap’s new album Ellipse. Good stuff.

Tue, 18 Aug 2009 07:17:00 -0400

Imogen Heap’s new album Ellipse. Good stuff.

How Did This Happen?

Tue, 04 Aug 2009 13:38:33 -0400

How Did This Happen?:

This is the last entry in my collection of NY Times apologies. There is no way to possibly top this: The NY Times apologizes at great length about the mistakes it made in Walter Cronkite’s obituary and another article about him. It’s pretty open and names several people who, for a variety of reasons, did not do their jobs properly.

But what is worse are the comments; they’re simply brutal. And I don’t feel that most of them are justified. Sure, making seven or eight mistakes in one story is pretty bad, especially for a newspaper with the NY Times’ reputation. But the mistakes are really rather minor, and they’re obviously embarrassed, so I don’t get what all the outrage is about. Especially because these are comments on the apology, not the original story. I wonder how many readers would have been able to spot those mistakes, or took away wrong information from the article.

Bobby McFerrin demonstrates the power of the pentatonic scale,...

Fri, 31 Jul 2009 14:36:11 -0400

Bobby McFerrin demonstrates the power of the pentatonic scale, using audience participation. He plays the audience like an instrument. Fantastic.

Moving from github to bitbucket

Sun, 26 Jul 2009 17:33:00 -0400

Open Source project hosting has taken a huge step forward over the last year or so. This was facilitated by the emergence of a new breed of version control systems, which are distributed rather than server-based (like CVS and SVN). There is a whole list of these systems now, but the ones that get the most mindshare are git and Mercurial (hg).

Collaboration in open source always used to be complicated and bureaucratic. While there is a lot of talk about how anybody can fix bugs in open source, etc., the practice was quite different. Most projects required you to prove yourself by submitting patches to the maintainers and earn their respect before you would be granted commit access to the CVS or SVN server. That’s understandable, they didn’t want you to submit bad or even malicious code; but it also meant that few people ever fixed bugs they found, because they just didn’t want to put in all that extra work. Many open source projects have also died a slow death by abandonment because there was no easy way for others to take over.

Distributed version control systems have made this much easier, and hosting services like github and bitbucket have added some simple but effective tools. Now, it’s easy to clone a repository, make changes, and then send a pull request to the maintainer. If the maintainer likes the code, he or she can incorporate it into the “official” version. If not, you can maintain your forked version, and the users can decide which one they prefer. It’s even possible to take changes from several people’s forks and combine them into a new version. That makes it possible to effectively take over a project, even if the original repository has been completely abandoned.

I originally signed up with github, and I really liked it. The only real downside was their use of git. Now git is without doubt the most flexible and powerful distributed version control system, but it has a lot of features I simply don’t need. I prefer using hg for its more streamlined workflow and better integration in eclipse. It’s also less of a hassle to set up in Windows and easier to explain to my students.

Perhaps bitbucket is just ripping off github, the two are really quite similar. But what made me switch was a simple limitation: Both github and bitbucket allow up to five private repositories on their lowest paid tier, but github limits the number of collaborators on these to only one. That’s just too limiting for me, and severely limits its usefulness.

Both provide cname support, which lets you set up a (sub-)domain to point to your github or bitbucket stuff. On github, you can create HTML pages, and they even include an elaborate way of creating static websites from configuration files that are hosted there. A cname domain points to your pages rather than your profile, so you can effectively use it to host your website there - but that’s completely useless for me. bitbucket simply points your domain to your profile page, and even changes your repositories’ clone URLs to use the domain. That’s a nice little touch and makes a lot more sense to me.

Maybe I’m making too much of this, but it seems that github is going the route of more complexity for little practical benefit - just like git. I prefer the balance of complexity and power that Mercurial (and, so far, bitbucket) provide.