3 ms·

Please, please think this through on a case-by-case basis. Adopting some poorly-maintained FOSS solution just because it's free instead of paying $10/mo isn't a

Please, please think this through on a case-by-case basis. Adopting some poorly-maintained FOSS solution just because it's free instead of paying $10/mo isn't always the right choice. It creates a lot of tech debt that you or someone else has to pay down later, often at a MUCH higher cost. Something like postman, for example, is easily a multi-month or year project for a big team of devs to create/maintain. I've been using the free version for years and it's saved an amazing amount of time. I'd use the paid version if I had a bigger team. Something like Cloudflare is a whole different scale altogether. Their free tier is a loss-leader for their enterprise plan, but even the free tier is better than most paid CDNs out there. And good luck trying to replicate that on a FOSS stack and shared infrastructure. If you're worried about secret billing, either don't provide your billing info to begin with, cancel trial subscriptions as soon as you start them (usually they won't expire until the end), or use a virtual CC# service like privacy.com.

allarm 5y ago
> Adopting some poorly-maintained FOSS solution Well, don’t adopt a poorly-maintained solution, adopt a well-maintained one. Simple.
- unknown_error 5y ago
  That's why it's a case-by-case thing. If there IS a well-maintained, well-documented etc. FOSS solution, of course use it (and contribute back!). But npm has the popularity/quality/maintenance metrics for a reason; a lot of smaller FOSS projects become abandonware once their original use case is dealt with, or developers turn over, or whatever. There's a lot of smaller packages/solutions/projects that just don't see much love anymore. I'm not against using FOSS or commercial. Just think there should be a case-by-case cost-benefit analysis. A $10/mo service should be evaluated differently than a $1000/mo stack, and labor value should be compared against both. Case in point, my boss recently asked me to build a new project that wasn't a part of our regular stack in a somewhat emerging field (web mapping). I did a brief market analysis and let him know that different vendors could probably get it done in about 2 weeks at about $2000. He said that was too much, I warned him that building it in-house would take several months and be very buggy because we have no specialists in that field on our team. He said to use FOSS software to do it, so we did... it ended up taking about 9 months and a few hundred hours of dev time. Now it's built, but with minimal features and lots of bugs. And nobody else can maintain it. The project was in turn open sourced, because we built it from open source software and wanted to give back, but honestly it's shit code. And we're considering abandoning it already due to the high overhead. But you know, someone else is going to run into the same situation at some point, and maybe find it and fork it because it's open source... then probably go through the same rigamarole... well, now we have a proper abandonware factory. shrug At the same time, I use (and thoroughly appreciate) redis, apache, varnish, chromium, brew, react, etc. Some of those are vendor-supported, some aren't. It's just a mixed bag. I'm just saying that choosing something just because it's FOSS so you save $10/mo is a lousy reason. There's a lot of great commercial software out there, just like there's a lot of great FOSS software out there.

allarm 5y ago

> Adopting some poorly-maintained FOSS solution Well, don’t adopt a poorly-maintained solution, adopt a well-maintained one. Simple.

unknown_error 5y ago
That's why it's a case-by-case thing. If there IS a well-maintained, well-documented etc. FOSS solution, of course use it (and contribute back!). But npm has the popularity/quality/maintenance metrics for a reason; a lot of smaller FOSS projects become abandonware once their original use case is dealt with, or developers turn over, or whatever. There's a lot of smaller packages/solutions/projects that just don't see much love anymore. I'm not against using FOSS or commercial. Just think there should be a case-by-case cost-benefit analysis. A $10/mo service should be evaluated differently than a $1000/mo stack, and labor value should be compared against both. Case in point, my boss recently asked me to build a new project that wasn't a part of our regular stack in a somewhat emerging field (web mapping). I did a brief market analysis and let him know that different vendors could probably get it done in about 2 weeks at about $2000. He said that was too much, I warned him that building it in-house would take several months and be very buggy because we have no specialists in that field on our team. He said to use FOSS software to do it, so we did... it ended up taking about 9 months and a few hundred hours of dev time. Now it's built, but with minimal features and lots of bugs. And nobody else can maintain it. The project was in turn open sourced, because we built it from open source software and wanted to give back, but honestly it's shit code. And we're considering abandoning it already due to the high overhead. But you know, someone else is going to run into the same situation at some point, and maybe find it and fork it because it's open source... then probably go through the same rigamarole... well, now we have a proper abandonware factory. shrug At the same time, I use (and thoroughly appreciate) redis, apache, varnish, chromium, brew, react, etc. Some of those are vendor-supported, some aren't. It's just a mixed bag. I'm just saying that choosing something just because it's FOSS so you save $10/mo is a lousy reason. There's a lot of great commercial software out there, just like there's a lot of great FOSS software out there.