5 ms·
This article hinges on the "productivity is subjective" argument, which isn't true in all cases. There are objective measures for certain types of productivity
by MrPowers 6y ago
This article hinges on the "productivity is subjective" argument, which isn't true in all cases. There are objective measures for certain types of productivity: library downloads, Stackoverflow points, blog views, GitHub stars.
I see developers adding high value features to the Apache Spark codebase that's used by hundreds of thousands of people. The top Apache Spark Stackoverflow question responder has hundreds of thousands of points (their answers are always the best).
Some folks build libraries that get thousands of downloads a day.
There are even popularity gauges that can be used for internal projects. Who is building the code that everyone is using and who is building stuff that gets thrown out.
This article also makes the fallacy of equating productivity with working hours. Working hours + productivity are correlated (top Spark devs seem to work constantly), but I also see people that work really hard and can't get traction for any of their open source / internal projects.
The power law is at play in software. A relatively small number of developers write the code that gets used the most. Yea, that might not be true for a given web app, but it's definitely the case for software overall.
- amelius 6y ago> The power law is at play in software. A relatively small number of developers write the code that gets used the most. Yes, but there is another law: if X doesn't write a certain library with basic functionality, then Y will. So your law translates to: we all use the same libraries for our basic needs. The question is, does it mean anything?
- Scarblac 6y agoEven so, it's often not the developer who gets to decide who writes what.
- eptcyka 6y agoI think there are far more truly productive developers than there are people with popular blogs, many a github star or SO points. Of course, these are indicators of productivity, but I feel these metrics are too specific.
- high_byte 6y ago> This article hinges on the "productivity is subjective" argument, which isn't true in all cases. There are objective measures for certain types of productivity: library downloads, Stackoverflow points, blog views, GitHub stars. Lines of Code is not a measurement, why would lib downloads be? Remember the time some 30-line JS lib who just does "split" or something like that was removed and it broke the Internet? heavily relied on - yes (ie heavily downloaded), highly productive? doubtful. You are referring popularity, which sure is part of the sum output of you as a person, but this article is meant to reference purely development productivity. Marketing or whatever is beyond scope. Maybe next post! (popularity gauge != productivity gauge)