Sometimes you read an article and it gets you giggling. No you don't? Okay, sometimes I read an article and it gets me giggling.

I have a history with the press, and it left me with two distinct beliefs. First, look for the bias in the article - it's there, whether it's a political article or a tech article, the writer has opinions and editors seem to have lost the ability to filter them, even when they're "warm fuzzy" opinions rather than fact-based ones. Second, don't assume that just because a publication was willing to publish this person they know anywhere near as much as you. This is particularly true in the weekly news-meets-tech space, but is true across the board. I know bloggers who've left more intelligence imprinted in their keyboards from pressing keys than some "published" writers I know of will ever have.

With all of that said, this week's laugh-fest came from reading this entertaining little piece in InfoWorld - Ruby hailed as economic solution, offering smaller investment and less risk. First off, if your title has to be that long, you missed the boat before you wrote the intro, but that's beside the point, we're here to talk tech.

My bias: I like ruby, I've used it, I like it with rails and without, I think it's a great extension of scripting languages. But it does indeed have performance problems and contrary to the FUD that its proponents put out, it is indeed a scripting language. Its performance problems are both legion and endemic. Now you know where I'm coming from and you don't have to look for my bias ;-).

My favorite quote - which didn't even get questioned by the interviewer even with the heaps of contrary evidence out there - is this one: "Really, the way you scale Rails is just like you scale any other Web app," Yeah, if you multiply other web app scaling requirements by a factor of ten or so.

The person interviewed goes on to discuss tests with "responses per second" as the gauge. What is defined as a response for purposes of this test? What size were the responses? How much processing was done to generate this response? The interviewer didn't answer any of these questions. It should be clear to even the most network-unaware developer that returning "hello World" in HTML format doesn't test the language at all, merely the bandwidth of the server. The interviewer didn't even ask if the test machines were the same exact config for these tests.

And then it goes on and on... Ruby is fast in "real-life web benchmarks" the author says. Oh Really? So what exactly is a "real-life web benchmark"? Is it real life, or a benchmark? Which benchmark is that, if it is one, so we can find out for ourselves?

Note, this is not just me talking. F5 believes in sharing the tests it uses to publish results, check out our testing information (note that anyone can get results, but you must be registered on DevCentral to download configurations and test specific).

Merb - the framework that the interviewee was selling - sounds great, and I'd love to try it out... But this article offered no useful information at all. Not a shred.

But it was great press for the Merb crowd, I suppose.

And it gave me some full-belly laughs, which is good for the soul, so I guess it serves a purpose.

If you want useful information on scaling ruby, check out the content here on DevCentral that is provided by Joyent, a ruby-based cloud provider that happens to be a customer also. Here's the interview with Jason Hoffman, their CTO to get you started, but a search of DevCentral on Joyent will turn up more conversation and documents. Since Joyent really has scaled ruby applications to handle millions of connections, we'll take their word as experts.

Meanwhile, I have to get back to other topics, just thought I'd share the funny article with you.

Don.

 

Share this post :