3 ms·
This has to be one of the most ignorant industry pieces I have read in recent memory. "Modern-day applications also don’t require large teams of engineers doin
by mberning 9y ago
This has to be one of the most ignorant industry pieces I have read in recent memory.
"Modern-day applications also don’t require large teams of engineers doing software development: They are user-customizable and can be built by anyone with basic programming skills."
Ok.
- tyingq 9y agoYou might enjoy his self written bio page: http://wadhwa.com/bio/ http://wadhwa.com/bio/ Needs more pictures of himself though.
- mythrwy 9y agoGeeze. I'd almost forgot how a lot of college really is. Thanks for the reminder.
- kevin_thibedeau 9y agoI don't think we can take him seriously as a distinguished fellow without at least one head shot of him facing right.
- statictype 9y agoYeah. The whole article went off the rails when I got there. Any idea what he is referring to?
- tdumitrescu 9y agoDidn't look at the article, but seeing this quote out of context I'd assume it's about cookie-cutter CRUD/CMS-based apps. Not that I would agree with that assessment, as even running those professionally and at scale requires experienced engineering staff.
- cookiecaper 9y agoToday, every app is written multiple times, usually at least twice for the two major smartphone platforms (sometimes a third or fourth time if they're trying to support Windows Phone, Blackberry, Firefox OS, or some other upstart phone platform) and at least once for the browser. It requires at least 3x the labor that was required during the relatively-brief period when everyone was cool with a browser-based web app replacing a desktop app and expected their mobile web browser to learn to read the internet instead of expecting developers to write a special version for their phone. Additional effort is also being split through the complexity prism of cloud server administration, CDNs, and devops. Everyone believes they need a high-availability cluster for everything these days. We've also moved away from the trend of straightforward RDBMS-backed applications and gone to the experimental realm of key-value stores, document databases, and trying to figure out some XPath-type thing for raw JSON structures. The author's assertion would've been more true about 10 years ago, but I don't think it was ever really true for anyone who wanted something more than a simple blog, which have been readily available off-the-shelf at least since LiveJournal became a thing in the 90s, so no real change there.