215 karma · joined April 26, 2016
Give it another chance if you want to, I’m always happy to help.
- At first (iPad2 & 3 years), for those who already had remote machines of some kind, the goal was to simply have a decent terminal on mobile devices to connect to them. The appeal was very clear and I think for a lot of people it offered a different context to do stuff. It still does.
- Things start to grow with iPad/iPhone, Apple was heavily investing on the hardware, and that brought another wave of users. The goal changed, it was not enough to "connect" anymore, but to do real work from these devices, for hours. The talk was to replace your laptop with an iPad, may sound familiar. We thought this wave was going to be big... but it never really materialized. Blame it on lack of software? The OS? M1? ...
- And so I think we are now on a third generation. I don't think the goal is to replace anything anymore. I see the next wave as a mix of the previous two. I want to work from anywhere, access my tools from wherever I am, access the code for my projects, pushing changes fast, etc. And it may be in my desktop computer, or in my coworker's laptop, on my iPad or my phone. Codespaces or in our case Blink Build are different approaches to do that. It is still getting started.
I think we are entering a world similar to what Rob Pike mentioned in this interview (https://usesthis.com/interviews/rob.pike/). The idea is that "state" is not something that you carry all day with you, but it is provided to you when you need it, connecting applications and dev environments from anywhere. That "state" may not necessarily be the Cloud, it will be computers in your office, or powerful remote environments somewhere, etc...
So yeah, if you were expecting Xcode, IntelliJ and XXX coming to an iPad and replacing your laptop, that may already be behind us and maybe even killed on phase 2. And although these apps may not play well in this new environment, it is worth questioning if it is more fault of the environment where we are headed or of the apps themselves. I want to think that Codespaces, Fleet, GitPod, Build, and your multiple machines wherever they are, will all play a part on that.
IMO, the best part of Apple adding Passkeys support over WebAuthn is that we are finally able to use Secure Keys as well across all devices. If this takes off, I wonder if it would somehow end up replacing FIDO2 for Secure Keys scenarios as wel.
Hope this solved your question!
I even had to go one night to Grand Central's Apple Store and buy an external BT keyboard to run some tests because I realized the Simulator wasn't working properly with the modifier keys. And fortunately I was right, I could make modifier keys work, and it made sense to continue working on Blink.
When I moved to the iPad-only I thought this would be an issue. Reality is I cannot code without WiFi anyway (checking documentation, doing research, etc...). I understand that may not be the case for everyone though, but coding without WiFi may actually be the exception.
And for spotty-low speed, that's where Mosh shines. I've even worked on spotty 2.5G tethering from a remote place.
It was done by one of our collaborators (advantages of open source), and it allows for multiple terminals just as you mention :)
Having a browser within the app has also been suggested, yo transform it in a nice dev tool instead of a terminal. We are still thinking what this could really be in the future,so any feedback and forward thinking is really appreciated!
Blink also has this problem, but Mosh is able to recover gracefully.
Blink also supports Cursor keys and remap of those (by default Cmd + Arrow).
My plan is to have some kind of "special keys maps", like iTerm does, to emulate any other key through defined combinations. That I guess should cover the rest of the special keys you might need.
Thanks for the heads up about the delay. I have only see that problem once with the iPad when I had another combination, do you have any other combination like caps as ESC? I definitely haven't seen that with other keys, and I'm really picky about my Cmd as Ctrl on Emacs :D
In the case of VLC they had to relicense, so a process very similar to this one.
It proved to be fast enough, and it is still faster than anything else. I'm considering if building on top of HyperTerm at some point would be doable too :)