I'm a staunch believer that watches should be small enough to never annoy, always show the time, and never have to be charged but I think this finally does all those things well enough that it's a legitimate option.
What is that warranty though? 30 days is pretty rough for a new and untested product. It's definitely enough to make me hold off for a year just in case.
I hear you, but the first new Pebble is getting good reviews. It's been out for months and I've not heard anything bad about it. While the upcoming models are a bit less derivative of past watches, they are still pretty well-grounded in prior Pebbles. The software is the same as the P2D, and the hardware is pretty similar to the prior rect/round Pebbles.
I believe things could go wrong, but I'm not sure what sort of latent errors would make sense to worry about. The battery life dropping precipitously? (Why would it do this?) Sensors breaking within a couple months? (Again, what would lead to this?).
I'm curious to know what you are concerned about. I agree that a year of track record would be great, but IMO a Pebble will be such a big upgrade over pretty much anything else out there (AWs etc. that need charging all the damn time), that I'd rather not wait a year for the additional data to come in.
Yikes, that sounds awful. I know that was an issue on the old Pebbles, and the ones that are out have some old hardware. The Pebbles that aren't out yet don't share any parts, so presumably this won't happen to them.
I would love to see some benchmarks of unison somewhere on their website. I find knowing there performance characteristics helps a lot with understanding the use cases for a new language.
Even just a really rough "here our requests per second compared to Django, express.js and asp.net"
Would be great to get a rough read on where it sits among other choices for web stuff.
More generally, I do hope this goes well for unison, the ideas being explored are certainly fascinating.
I just hope it one day gets a runtime/target that's more applicable to non web stuff. I find it much easier to justify using a weird language for a little CLI tool then for a large web project.
In years of using eBay, have never had an issue with it. Sad that that's a high praise these days, but it is. eBay is fast, it works damn well, and always has.
As a counter point, React's poster children, in messenger, Facebook and Instagram. Have all been plagued with UI bugs for the entire time I've used them.
Obviously those aren't wholly comparable, but I do think it's worth taking note of the actual outcomes we have when tools are used at real scale.
As someone who has lightly used Haskell and quite heavily used ocaml:
- Pragmatism: The ocaml community has a stronger focus on practical real projects. This is shown by the kind of packages available on the ecosystem and the way those packages are presented. (A number of Haskell packages I've tried to use often seen to be primarily intellectual pursuit with little documentation on actually using them).
- Simplicity: Haskell does have some amazing features, but it has so many different ways to do things, so many compiler flags, code can look vastly different from one codebase to another. It's kind of the c++ of FP. Ocaml is definitely more cohesive in its approach.
Tooling: last I used Haskell the whole ecosystem was pretty rough, with two package managers and a lot of complexity in just setting up a project. Ocaml is not at the level of rust or other modern languages, but is definitely a stop above I'd say.
What is that warranty though? 30 days is pretty rough for a new and untested product. It's definitely enough to make me hold off for a year just in case.