Does Software Need a Siskel and Ebert? 169
theodp writes "Over at Scripting News, Dave Winer laments the lack of serious software reviews in the NY Times. That wasn't always the case, recalls Dave. 'When they started doing software reviews in the early '80s it was with the usual Times flair,' says Winer. 'But somewhere along the line they stopped taking tech seriously. It's as if they would only review Saturday morning television shows. How could television like The Sopranos or Breaking Bad take root in the culture if there was no criticism that discussed it? Yet that's where we are today with software.' So, does software need a Siskel and Ebert (or A.O. Scott and Manohla Dargis for you highfalutin NYT readers!)?"
We need reliable reviews (Score:5, Insightful)
No (Score:3, Insightful)
No. Siskel and Ebert rocked don't get me wrong, but we have a thing now called the Internet and google which can pretty well give you any info you want on most software out there. Anything in a newspaper or magazine is going to be influenced by $$$ anyway while the Internet is typically pretty damn raw
Re:Ebert already rated software (Score:5, Insightful)
Re:We need reliable reviews (Score:4, Insightful)
I guess if your job is to talk about phones, and all the phones are pretty similar, it's very easy to develop strong opinions about trivial details. Oblig XKCD [xkcd.com]
Software "previews" are way longer than a movie... (Score:5, Insightful)
Part of the problem is that a critic can sit down for two hours watch a movie and write a meaningful review. This is not possible when it comes to software.
Let me use a real life example: I was an early proponent of Java since my first few interactions with it in 1994 were positive. Only when I was is deep in the bowels of the beast did I start to see the problems: flawed parameter passing model, the "everything has to be an object" religion (which ironically is violated by built in data types), the "you must write a preamble bigger than COBOL's to have a well designed piece of code", the horrible graphics library that if first shipped with, etc.
After that I realized that maybe moving to Java is not such a good idea after all. I think the popularity of C#/Haskell/Scala/C++11/Python are a result of this realization.
Re:Software is too plentiful (Score:4, Insightful)
My sentiments exactly. Whether you're talking software written by multi-billion dollar companies such as Oracle or SAP, to smaller companies or homegrown software, the current state of software is abysmal.
"Throw more RAM at it!" is the usual response, as if that solves the underlying problem. Worse, you can have identical machines and get different results when installing the same piece of software.
The biggest problem is no one is held accountable for this nonsense. Unlike building a bridge where you can check to see if the designers did their job, the engineers did their job and the construction folks did their job, there is nothing similar in software. At best, you have to wait for a patch which might, maybe, possibly, fix some issues, but then again, maybe not.
Re:Software is too plentiful (Score:4, Insightful)
In days of old, before the Black Ships came and the secret of hose gartering that never ravels was lost and forgotten, Niklaus Wirth figured this out [inf.ethz.ch] and bequeathed us Wirth's Law. [wikipedia.org]
Back when the building RSX-11 executables larger than one MB that would consistently execute in real time required manually mapping memory for the taskbuilder step, software engineers had to write rockin' code just to survive in the field. We were all computer scientists by necessity. Today, though, the barrier is pretty low; just slap together a bunch of Java modules some anonymous 13-year-old wrote in a GUI and call it programming.