Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Popularity-driven development

ferrante.pl

Post reply on HN
Print view View original

11–20 of 36 posts

Re: Popularity-driven development

#11

Post by CJefferson »

cauterized wrote: post #9

Indeed, popularity does not directly correlate with quality (which is not to say that popular libraries and tools are never high quality.) That said, the reasons most projects fail have less to do with the imperfections of the tools used than with the approach taken to the project as a whole. And especially if you're running a business there are good reasons to choose tools that are reasonably well known and widely u…

I agree entirely. Long ago, I got some amazing advice, which has helped me to this day.

"When you have n choices, and you aren't positive which to choose, choose the more popular one".

I used to think this advice was about "the more popular is better". Now I don't think that's true. But the more popular is (in the case of software for example) more likely to be tested on more OSes, or have questions with good answers on stackexchange, have people you can employ to work on it, and still be functioning in 5 years time.

The same arguments apply to hardware too -- much easier to convince Android developers to fix their app if there is a bug on the Galaxy Note (for example), than some obscure handset.

Re: Popularity-driven development

#12

Post by woah »

cauterized wrote: post #9

Indeed, popularity does not directly correlate with quality (which is not to say that popular libraries and tools are never high quality.) That said, the reasons most projects fail have less to do with the imperfections of the tools used than with the approach taken to the project as a whole. And especially if you're running a business there are good reasons to choose tools that are reasonably well known and widely u…

Something that is sometimes seen as a "marketing" type thing is documentation. However, good documentation is a clear indicator of a better project, if only because it makes it qualitatively easier to implement. I'm going to choose the project with the clear, attractive, well written guide and api docs over the project where I have to dig through pull requests to find method names.

Worse are the projects that actually expect you to look at the tests as the only form of docs. This indicates that the author does not care about whether you have an easy time using the project, and that it was only written for their own amusement.

A well done project website shows that the author wants you to have an easy time working with the code, and is "putting out a welcome mat".

Compare mercury: https://github.com/Raynos/mercury with vue: http://vuejs.org/.

mercury is likely the "better" framework, with it's reactive approach, but in vue, there is attention to detail and documentation through and through. So vue is inevitably what I recommend to backend developers or others who just want to get something done.

Re: Popularity-driven development

#13

Post by semperfaux »

qznc wrote: post #6
comboy wrote: post #2

Especially in ruby world, it's amazing how big impact marketing has on open source projects. It's nothing new that when you prepare a nice webpage with some pictures and logo your project is more likely to gain some traction. But mentioned element of social media and virality makes sometimes seemingly random projects to become popular. It would seem that when releasing an open source project, your webdesign and marke…

So instead of discussing "Worse is better" vs "Worse is worse", we should talk about "Presentation is better".

"Better presentation is more effective at attracting attention" would, I think, be the only reasonable conclusion here (and then only because it seems axiomatic).

In that context, it would be interest to talk to someone like kalzumeus about testing developer response to various methods of presentation for open source projects, not only to see what works in that regard, but to see whether contributors and adopters respond to the same sort of cues you'd expect to encourage consumer behavior.

Re: Popularity-driven development

#14

Post by jnbiche »

woah wrote: post #12
cauterized wrote: post #9

Indeed, popularity does not directly correlate with quality (which is not to say that popular libraries and tools are never high quality.) That said, the reasons most projects fail have less to do with the imperfections of the tools used than with the approach taken to the project as a whole. And especially if you're running a business there are good reasons to choose tools that are reasonably well known and widely u…

Something that is sometimes seen as a "marketing" type thing is documentation. However, good documentation is a clear indicator of a better project, if only because it makes it qualitatively easier to implement. I'm going to choose the project with the clear, attractive, well written guide and api docs over the project where I have to dig through pull requests to find method names. Worse are the projects that actuall…

I don't use Vue.js (I use React), but I agree that the project is extremely well-documented and beginner-friendly.

The project also has 3400 stars on Github, so the lead dev is gaining some significant popularity. Furthermore, a number of for-profit businesses are almost certainly using the framework in production, as a key part of their websites (Optimizely is cited on their web site as using it in production, but I'm positive there are others).

And yet, for all his efforts, the developer is given $5.00 a month on his prominently displayed Gratispay account: https://gratipay.com/yyx990803/

$5.00. A single Starbucks cappuccino. A month.

When I see things like this, it makes me despair of ever making a decent living doing open source software (I have some semi-popular projects, but nowhere near the level of Vue.js).

In the end, it also makes me wonder if open source development isn't devaluing the software profession in terms of compensation and prestige. If seems like if people don't pay for something, they just don't respect it, regardless of how useful, well-documented, and well-designed it may be.

I love open source, but every time I see a super popular project with only a few dollars of total donations, or hear the lead dev of a open source project used by massive companies around the world tell me he's barely making a living, part of me dies inside.

Re: Popularity-driven development

#15

Post by kyllo »

comboy wrote: post #2

Especially in ruby world, it's amazing how big impact marketing has on open source projects. It's nothing new that when you prepare a nice webpage with some pictures and logo your project is more likely to gain some traction. But mentioned element of social media and virality makes sometimes seemingly random projects to become popular. It would seem that when releasing an open source project, your webdesign and marke…

Marketing shows that a project has some money behind it, which is a good sign it's going to be supported for a while, which allays the biggest fear most people and organizations have when selecting an open source tool: "is it abandonware?"

Re: Popularity-driven development

#16

Post by drapper »

Popularity might not be the best indicator of quality, but it's a pretty good indicator of longevity of the project - and this is, I think, more important, in the long term at least.

All the technical superiority of an unpopular library will amount to nothing if couple of months down the road the author will not be interested in fixing some obscure, but important bug, nor there will be anyone to take on the maintenance from him/her.

Re: Popularity-driven development

#17

Post by awinder »

drapper wrote: post #16

Popularity might not be the best indicator of quality, but it's a pretty good indicator of longevity of the project - and this is, I think, more important, in the long term at least. All the technical superiority of an unpopular library will amount to nothing if couple of months down the road the author will not be interested in fixing some obscure, but important bug, nor there will be anyone to take on the maintenan…

Importantly, like you said, it's also the measure of some level of quality. It's incredibly, incredibly rare that a tech project becomes popular without any level of quality. A lot of people could do much worse than that combination (some quality + longevity)

Re: Popularity-driven development

#18

Post by dragonwriter »

awinder wrote: post #17
drapper wrote: post #16

Popularity might not be the best indicator of quality, but it's a pretty good indicator of longevity of the project - and this is, I think, more important, in the long term at least. All the technical superiority of an unpopular library will amount to nothing if couple of months down the road the author will not be interested in fixing some obscure, but important bug, nor there will be anyone to take on the maintenan…

Importantly, like you said, it's also the measure of some level of quality. It's incredibly, incredibly rare that a tech project becomes popular without any level of quality. A lot of people could do much worse than that combination (some quality + longevity)

Nothing exists without some (possibly very low) level of quality, but by any measure of quality other than popularity itself, popular things may fall anywhere in the quality spectrum.

Re: Popularity-driven development

#19

Post by hippich »

CJefferson wrote: post #11
cauterized wrote: post #9

Indeed, popularity does not directly correlate with quality (which is not to say that popular libraries and tools are never high quality.) That said, the reasons most projects fail have less to do with the imperfections of the tools used than with the approach taken to the project as a whole. And especially if you're running a business there are good reasons to choose tools that are reasonably well known and widely u…

I agree entirely. Long ago, I got some amazing advice, which has helped me to this day. "When you have n choices, and you aren't positive which to choose, choose the more popular one". I used to think this advice was about "the more popular is better". Now I don't think that's true. But the more popular is (in the case of software for example) more likely to be tested on more OSes, or have questions with good answers…

be careful with it tho. it also depends on the reasons why project is popular. for example, wordpress is popular is because it is so easy to setup and bring to life for person with little to no technical knowledge. but other side of the medal is that code you see contributed by community often will be of very low quality. Same applies to javascript frameworks, and I bet, to hardware too (although I don't have example for hardware of top of my head)

Re: Popularity-driven development

#20

Post by mikegerwitz »

drapper wrote: post #16

Popularity might not be the best indicator of quality, but it's a pretty good indicator of longevity of the project - and this is, I think, more important, in the long term at least. All the technical superiority of an unpopular library will amount to nothing if couple of months down the road the author will not be interested in fixing some obscure, but important bug, nor there will be anyone to take on the maintenan…

> Popularity might not be the best indicator of quality, but it's a pretty good indicator of longevity of the project

Certain projects have either a niche audience, are less known, less easily groked by users, etc.

Post reply on HN