Rethinking devtools: Escaping the Cloud and SaaS trap
9 comments
There were (and still are) a lot of advantages to being able to develop software on your own machine, disconnected from a network. Chiefly, that the only one who could break the setup was you, and if you improved your setup by making your build faster, etc., you got to reap the benefits.
Totally agree on this. There are things that belong in cloud, even a 3rd party one. Not everyone has a luxury of self-hosting. But keeping API secrets in any 3rd party cloud is just insane imho.
100% on this! It's sad to see why hardly any of the new devtools work without any outside network connectivity. Hope the era of network sandbox comes again
Working on it. It's hard to explain to devs who used to something how bad that something is. Especially when it once was awesome.
I wrote my own Postman alternative because I wanted something lower level that did both raw requests and raw responses without reliance on a GUI.
Something internal, or did you publish it?
What would you say is the most important stuff to you? Only simple API testing, or you gotta take care of the specing and documenting it as well.
It’s on github buried in a much larger personal project that I have not shown anybody.
I was once a JS developer for an eternity but now I do enterprise API management in a highly restricted environment. I just needed something reliable and postman was not always reliable.
I was once a JS developer for an eternity but now I do enterprise API management in a highly restricted environment. I just needed something reliable and postman was not always reliable.
Honestly, I think one of the main reasons these tools are cloud-based is that they're easier to monetize. Programmers are really into hacking, and it's hard to earn any real income from the tools you develop.
Yep, sounds like that to me too.
Pay-per-seat works well when no proper competition to go against it. What we built with http://voiden.md should be free forever, with monetization on plugins, but only the ones that introduce costs to the team.
The blessing was that the team was already profitable on another tool, and VC-independent, so nobody shoved some dumb design decisions down anyone's throat.
Pay-per-seat works well when no proper competition to go against it. What we built with http://voiden.md should be free forever, with monetization on plugins, but only the ones that introduce costs to the team.
The blessing was that the team was already profitable on another tool, and VC-independent, so nobody shoved some dumb design decisions down anyone's throat.
If not for a gazillion other reasons, a recent outage from Postman should remind us again why proper design decisions matter in devtools.
API workflows shouldn’t live in the cloud. A LOT of devtools shouldn’t live in the cloud. Some tools can’t work without it, all good. But a whole lot of them are there just because of some poor design choices.
Yet…
- No cloud? No outage.
- No cloud? No sync issues.
- No cloud? No data security gaps.
So why keep forcing it where it doesn’t belong?
And it’s not even just the cloud. Cloud is awesome for some stuff. SaaS-like designs are dragging devtools into the same trap.
As a matter of fact, devtools in general shouldn’t feel like SaaS platforms. Unless we’re going through app usage dashboards - tabs, and mouse actions are not built for devs.
Devtools should prioritize developer control, not just pay-per-seat subscriptions.
P.S. No need to send me Postman alternatives. I already work on building an awesome one, but chose not to promote it in the post.