Hacker News Reader

Comment by brudgers | Hacker News Reader

From Slate back in July: an interesting background piece about the campaign's approach to staffing for data analysis.

http://www.slate.com/articles/news_and_politics/victory_lab/...

What the debacle brings to mind is the consequences of putting "idea guys" in charge of software development when there is a hard deadline.

"Orca had been conceived by two men—Romney's Director of Voter Contact Dan Centinello and the campaign's Political Director Rich Beeson."

I imagine them saying, "We shall have an app," while thinking, "How hard can it be?" While the developers were rationalizing, "Well, a web app is technically an app...and a lot easier to complete than an iPhone app, an Android app, and the web app we will still need to build for the old people with a Windows desktop."

View on HN

A Web App is actually the right call. It was for volunteers doing GOTV in the field, not as much for people sitting at a desktop. And you would want to be able to issue updates quickly and carefully control access -- things that are much more difficult with an App Store app.

Using Windows was a mistake, that's for sure.

Good thing for Democrats they didn't know about the open source ushahidi platform.

How did you determine using Windows was a mistake from the information presented in the article ?

Cause it's a bias of his, likely.

What I get from the article is a classic tale of crappy IT management, the like of which I've seen and heard of several times.

"Using Windows was a mistake" is a truism.

Also, from the article, project failed. Failure is regularly equated with having made mistakes.

> "Using Windows was a mistake" is a truism.

No. No, it is not.

> Failure is regularly equated with having made mistakes.

And what mistake did Windows make here? Keep the baseless insults inside your head, please.

> "Using Windows was a mistake" is a truism

If you use or have used StackOverflow or Newegg, realize that you too are using Windows. Projects' success and failure is much more correlated with the folks running them than the technologies they use.

In the same vein, one could claim use of humans to write code was a mistake. After all, the project failed, this means they made a mistake, so obviously they should have used trained hamsters instead. Do you see where your logic failed you?

A: ungraceful crash under load

Sorry, a database-backed web app works just fine on a Microsoft stack, there are thousands of them out there and I say that as someone who vastly prefers unix but this is clearly a project management fiasco and architecture design fiasco, not a technology stack fundamentally unable to handle the demands.

I agree the MS stack works fine in the dev lab.

from the down votes apparently it doesn't work in the lab either!

No, just that you're plainly talking out of your ass.

Stack Overflow (you may have heard of them) runs a Windows stack with Linux used only for caching & logging (because it's "cheap").

http://blog.stackoverflow.com/2008/09/what-was-stack-overflo... http://highscalability.com/blog/2011/3/3/stack-overflow-arch...

Which also can happen with a terribly coded application running on Linux. Or with a terribly configured apache server.

This was a meatspace failure, more than anything else.

Why do you choose to throw out an insult at Windows? Is there anything about this story that points to Windows being at fault? How about you think before you flame next time, okay?

I have been a professional webmaster since 1995. I have worked for California's largest privately held webhosting company. I am an MCP in Windows server and have been the only guy on staff willing to support customers with IIS.

I base my statement partly on that experience.

Stackoverflow crashes daily

Don't take it personal. Inferior products need support, too. You will always be able to find work making IIS sites as long as MS is selling server software.

Looks like someone has a bit of an emotional investment on a tool, eh?

Nope, I just strongly dislike tangential and unfounded insults.

You can fail on any platform with bad implementation and/or bad management and/or bad planning. That's the lesson here.

(and at Twitter when they fail whaled for months on end with Rails - it wasn't Rails fault - it was the application of Rails in too high a volume deployment)

False equivocation BS.

Obviously, Dick Costolo is unfit to be the president. Will have to write it down in case he decides to run.

NGP-VAN, which is probably the most important company in the Democratic sphere, uses a Microsoft stack. And it didn't lead to a spate of post-election recriminations about how its tech fell apart.

Using the MS stack for a situation where you will have huge spikes of activity on an app with an expected lifespan of less than 24 hours is not best practice.

I'm an open source developer – working on Linux servers since 1997 or so, etc. etc. Please stop what you're doing: it's not very funny and it's not helpful, but it does give OSS people a bad reputation.

You can host plenty of high volume things on Microsoft's code – Azure, Bing, etc. have proven scale. We can argue about cost, ease of use, etc. but a decent team should have no trouble making a serious service. The problem here is that they didn't hire a decent team – from the sounds of it, it was the usual executive-level schmoozing and managerial incomplete you see throughout large organizations of every type. They could have used HN'a favorite stack and failed just as hard.

Hey, dogs can swim but that doesn't mean I use them to pull a barge up the river.

I will not shut up about the ground truth I and many other people know about Microsoft and their software.

The way you're doing it won't will anyone over. Open source has been winning where I work because we've demonstrated better results with less cost. I suggest you try to outperform rather than out-market.

> managerial incomplete

I'm guessing you wrote this post using autocompetence?

Sigh, yes. iOS auto-uncorrection to the rescue!

This comment reminds me of AL Gore lecturing Frank Zapa.

A poor craftsman blames their tools.

"A Web App is actually the right call."

Yes.

Calling it "an app" however was a bad call. From a practical standpoint, that is. From the standpoint of "an idea guy" and a "communications expert" it was of course a requirement.

What else do you call it? A "mobile website"? I don't think that's the real problem... the real problem is none of the users actually had an opportunity to use prior to it being absolutely mission critical.

"It's a website: http://orca.mittromney.com. It works on smartphones."

But so much can go wrong with an app, especially in the field. Give every team leader a fax machine, a phone, and some pens.

There's nothing wrong with web app over mobile app. Actually, recently here on HN was a number of articles about coming trend of switching from native apps to HTML5/web apps. It is a valid choice, especially when you need to communicate a lot of data, so you need network connection anyway.

Of course, that web app needs to be working - but the choice itself is completely fine. Given the circumstances, I would seriously consider going the same direction if I had to do an app like that.

Reply on news.ycombinator.com