As someone self-taught, and actively attempting to benefit from others known and shared experience, I don’t have anything more than anecdotal data to add. But compile enough anecdotes — well just anecdotes and you’re a medical study — enough anecdotes with metrics and you have a quasi-quantitative study.
I came to coding at all because my company could not afford 40,000 a month for developers and no results. Our best development experience came via a 19 year old, with one year of college cs. He built a front and back end that worked on mobile and all platforms in six weeks using ruby on rails, which he had never used before. The system was built in 2012. While not “pretty”, it worked and was extendable. A second coder with far more experience cut down its data usage for 4x the price the next year, but added little. A team with degrees, and lots of experience burned about $250,000 and produced nothing and left us less functional than before.
A self-taught young professional got us functional in like 5 weeks in Angular 1, but he has moved on to a full time gig at a big place, and now we are jammed again. My desire is a to totally rewrite soup to nuts, so we have a far more dynamic system that doesn’t just barely handle the data needed at the moment and barely handle the client “records” without real records. `
System issues have persisted since 2010, and no real solution has ever been suggested by any team or group we approached. Network Solutions team was an unqualified disaster, creating websites worse than the “looks circa 1998 and craigslist” one we first through up with power points as its visuals. (price $7000–14000 a month 0 usable pages or databases, used jumla against explicit instructions not to).
So, I wouldn’t say I have the data you want, but I have 6–7 years of unbelievable anecdotes. For better data, see the 2014-january 2015 Forbes series on “needing only full stack programmers”.
At my current skill level, I am could pretty much handle working in any framework. But I quickly butt-up against the frame limits. Since september 15, I have gone native only, and am now running into the holes in native and my own knowledge/accessible knowledge. Speech synthesis doesn’t seem like it’s ported to modern js. Wai standards make some approaches which are clearly more efficient “illicit”. The dom clearly has recursive loops in it, how are they handled? When is memory alterable. What fills the gig that can load up in memory while building? How does one limit that memory and still record a history, share instances? Should frames be used, as all big sites use them, to just shift between contexts and dash that memory usage. I have no idea what the answers to these questions might even look like, and I know no one who can answer them. So, even approaching another team of engineers seems like a crazy proposition.
I did have one meeting, where engineers insisted jQuery was the hallmark of sensibility, that meeting ended quickly. So, if you have data… please, put it up. If there is a venue to even address the poorly framed questions above, all business owners not of less than medium size (anyone pulling in less than like 20 million a year) needs answers. Since 2006 the loss of efficiency in those businesses and the resultant inability of those folks to hire and expand has cost this country an up to 1/3 bump in gdp. It’s the biggest loss of our intellectual capital that could possibly occur, and the current path for standards etc. does not bode well for its imminent correction.
To put it bluntly, and perhaps smugly/arrogantly — if I am having a tough time navigating path and discerning value in approach, then its a problem of huge size, which I still believe has a simple, straightforward solution, when all the politicking is removed.
Or I could just be clueless.
Not sure if this helps, Eric. I hope someone has this data for you and for all of us.