Experiment #4 – New interface in 1Password.com’s vault view

1P_Rob I think at this point we're at least pretty much on the same page about things. There's just two points I'd like to clarify:

we're going to disagree about what constitutes half-baked, and code at 1Password is always under active development.
It's fine to disagree, but I can at least offer a little context on why I used that term here. Definitely at the top of my mind is the huge drop in performance, with page loads taking several times longer than before. I think that alone is a hard blocker to calling something ready to ship. The other factor is that there's still a number of known issues being actively discussed in this thread, which is a pretty clear sign of instability, even if most of them don't really impact the way I use 1Password myself.

My personal rough definition of "stable" software is that it should be able to go some nontrivial amount of time (let's say a month or so) without requiring any changes that a typical user would notice. The old web UI has been meeting that criteria for some time now. Updates that improve performance or security are also perfectly fine from this position, while frequent bugfixes are a sign that something shipped too early, and UI changes (and to an extent, even new features) should, in my view, only be happening once or twice a year max. Feature updates once a year can be cool and exciting to check out all in one sitting, but tiny feature updates all the time are exhausting and make me not want to engage with any of the new things in the first place.

Basically, if things are still moving around frequently, I don't think you can really call that UI stable. I'm also aware that this is becoming an increasingly unpopular view among other developers and I'm unlikely to see a reversal of the rolling release trend, but I'm still not thrilled about it and at least make an effort to put my money where my mouth is by paying a premium for LTS versions of software whenever I have the option to do so. That really just comes down to a business decision, and as I recall from the days of 1Password 8 being released and users asking for a way to keep support for 7 going, a similar proposal was put forth and ultimately not taken up.

From a cold start, clicking edit in the extension, it took about 10 seconds for the item to show up for me in the new view in edit mode. Subsequent tests were much faster though, about 3.5 seconds each time.
I think I can offer a little context here as well regarding the way I use 1Password, or more specifically browsers in general. In particular, I make very heavy use of private browsing windows -- nearly all of the "browsing" I do happens in one, as does a significant portion of my login sessions on various sites. Pretty much the only things I open outside of a private window are tabs I want to be open 24/7. Even things like logging into bank accounts happens in those little ephemeral sessions. This upsets my bank (and plenty of other sites) to no end, thinking I'm logging in from a "new device" every time, but I much prefer keeping the persistence of sensitive data on my machine to an absolute minimum.

This means that when I'm editing an item in 1Password, there's a good chance that I'm doing so in a browsing session that's never seen my.1password.com before, and therefore wouldn't have anything cached from it. I'm almost always loading the UI completely from scratch. I also rarely have a reason to edit multiple items at once; I want this to be a quick in-and-out process.

Maybe there's a way to cache parts of the web app in the extension itself to poke through that session partitioning (which isn't really a privacy issue since the web app doesn't load any third-party resources), but that sounds pretty complicated, especially to cover a use case that probably isn't all that common. You could also work around it by having the extension always open the webapp in a non-private window, although that can quickly get messy if I've got a bunch of windows open, and I do like keeping related things closer together by always opening the webapp in the current window.

I realize this is incredibly out of scope for this particular project/discussion, but adding the ability to create/edit saved items directly within the browser extension itself sounds like it would perfectly address the bulk of my issues. I care a lot less about a page taking 10 seconds to load when it's not something I'm pulling up possibly multiple times a day, often right in the middle of another task. That would also be way less intertwined with some random user's specific usage patterns, and is something I'm sure would be appreciated by a much broader segment of the extension-using population. Now that I'm thinking about it, I have to imagine this is a feature that's already been requested before, in which case I'll enthusiastically add my vote to the tally.