Mastodon Feed: Posts

slightlyoff@toot.cafe ("Alex Russell") wrote:

In consulting with teams, this has come through in several ways. The most dispiriting of them have been teams that, when presented with evidence that their JS-heavy, unusable site isn't working for users, bleat out some half-hearted nonsense about how native apps are better for mobile anyway.

No? Almost never? And if you're going to give up on your own bloody profession, at least have a modicum of creativity about it? Or the guts to actually jump ship to native yourself?

slightlyoff@toot.cafe ("Alex Russell") wrote:

One of the most dispiriting trends of frontend's lost decade has been the capitulation: the idea that it's OK if mobile goes to native because native somehow has some superpower that HTML/CSS/JS lack.

It was always bullshit. It's more bullshit now than ever, and the only cases in which it isn't are due to Apple's abuse of it's monopoly on wealthy users (which, from a regulatory redress standpoint, at least has some positives in the US, UK, Japan, AU, etc.)

Reblogged by slightlyoff@toot.cafe ("Alex Russell"):
jcgregorio@mastodon.cc ("Joe Gregorio") wrote:

fribbledom ("muesli") wrote:

Good news, everyone. The days are getting longer again.

Reblogged by slightlyoff@toot.cafe ("Alex Russell"):
scrumble_eggs@mstdn.party ("Scrümbled Eggs :verified_red:") wrote:

Reblogged by bendingoutward:
shelby@fauxzen.com ("Shelby DeNike") wrote:

bcantrill ("Bryan Cantrill") wrote:

@igb @ahl @oxidecomputer If you signed up for this event before ~20:15 Pacific on 12/21, note that the original event was erroneously deleted; please signup with this updated link:

https://discord.com/invite/3jSyYpGP?event=1055335241411215383

(The original post has been updated to have the correct link.)

Can't wait to talk with @igb on Monday!

Reblogged by teller ("Siim Teller"):
ekp@mastodon.cloud ("Ellen Pao") wrote:

Attachments:

• 44f5f441d2de8647.jpeg (remote)
• 0d740f9182bb382b.jpeg (remote)

Reblogged by teller ("Siim Teller"):
BalticVC@mastodon.world wrote:

Reblogged by slightlyoff@toot.cafe ("Alex Russell"):
carnage4life@mas.to ("Dare Obasanjo") wrote:

Reblogged by slightlyoff@toot.cafe ("Alex Russell"):
braedon@mastodon.nz ("Braedon") wrote:

slightlyoff@toot.cafe ("Alex Russell") wrote:

Once upon a time I did the math, and for a relatively recent iOS & Android release cycle, in the trailing 60 days after the bits went to prod, iOS had upgraded more users *per day* than Android, despite having 1/7th the market share.

slightlyoff@toot.cafe ("Alex Russell") wrote:

In this data set, Android 9 + 10 (together) have more share than 12!

Heads should roll for an ecosystem fsckup that tremendous, but I guess as long as you get to predate problem gamblers through Play In-App Payments, life is good?

slightlyoff@toot.cafe ("Alex Russell") wrote:

Just got a copy of the Scientia Mobile Mobile Overview Report (MOVR)[1], and it's not the most important stat, but it's hilarious how I'm sure that Google's managers still think that releasing new versions of Android matters when, more than a year after Android 12's release, < 20% of users have it.

Like, stop? Just stop. It's not working. The whole model; totally broken, top to bottom.

[1]: https://www.scientiamobile.com/movr-mobile-overview-report/

Reblogged by slightlyoff@toot.cafe ("Alex Russell"):
benjaminwittes@thecooltable.wtf ("Benjamin Wittes") wrote:

Attachments:

• d25043e1439c4a18.png (remote)

Reblogged by kornel ("Kornel"):
craigsilverman wrote:

Attachments:

• f39b16e42dc1aeef.png

slightlyoff@toot.cafe ("Alex Russell") wrote:

In multiple threads over on birdsite, I've had multiple tell me (when faced with actual data and traces) that it didn't matter if modern web apps were terrible because "everyone uses the native app for that"...and IDK, man, but like, isn't the point of all this to avoid sending users to native apps?

And if "modern frontend" is just a fifth-column for native, maybe it *should* be the enemy?

Reblogged by slightlyoff@toot.cafe ("Alex Russell"):
ntx@hachyderm.io ("Nodir 🇺🇦") wrote:

Reblogged by slightlyoff@toot.cafe ("Alex Russell"):
darius@friend.camp ("Darius Kazemi") wrote:

bcantrill ("Bryan Cantrill") wrote:

@igb became an instant folk hero when he called out the world's wealthiest man, taking him to task for his grievous misperceptions on how software systems actually work:

https://twitter.com/iamraisini/status/1605538033612177408

While brief, this exchange hit on some deep issues in software engineering; @ahl and I are thrilled to announce that @igb will be joining us for our @oxidecomputer and Friends on Monday. Join us as we break it down with Ian Brown:

https://discord.gg/shdPpXSu?event=1054558914982707220

Recorded, as always:

https://feeds.transistor.fm/oxide-and-friends

Reblogged by jeffsonstein ("Jeff Sonstein"):
dadrummond@mastodon.online ("D. Allan Drummond") wrote:

Attachments:

• 0ad98aeb589365b8.png (remote)

Reblogged by jeffsonstein ("Jeff Sonstein"):
UncleDuke1969@mstdn.party ("Uncle Duke :verified:") wrote:

jeffsonstein ("Jeff Sonstein") wrote:

I Never Loved a Man (The Way I Love You) (2018 Mono Remaster) by Aretha Franklin

https://pandora.app.link/l7WHrWYaXvb

Attachments:

• fa306dc764536cf9.png

Reblogged by slightlyoff@toot.cafe ("Alex Russell"):
allafarce@hachyderm.io ("Dave Guarino") wrote:

slightlyoff@toot.cafe ("Alex Russell") wrote:

Obviously, doing so would shrink the market for inappropriate uses of technology that, in the main, isn't working.

I can see why they might not feel moved to make such a case. But we need to apply a more metrics-based, analytical lens to these claims, even if the thinkfluencers and salesmen won't.

jeffsonstein ("Jeff Sonstein") wrote:

Leather and Lace (2016 Remaster) by Stevie Nicks & Don Henley

https://pandora.app.link/ik0pvVX9Wvb

Attachments:

• 898e610e9a877a14.png

slightlyoff@toot.cafe ("Alex Russell") wrote:

Basically, you want to divide your cumulative costs by the number of trips through the interaction loop per session.

A theoretical case for this technology that falls out of a tiny model like this, and apologists for client-side bloat (which, at this point, seems to be most of the frontend world?) could be putting forward metrics with interaction depth amortization too...but they haven't.

slightlyoff@toot.cafe ("Alex Russell") wrote:

Now, it's possible for us to come up with metrics that help illuminate how we can architect for better UX all around, and such metrics would start with session depth.

E.g., here's a tiny mental model I've been sketching on whiteboards for a while to anyone who will listen:

Attachments:

• The shallower your sessions, the less client complexity you can afford. (remote)

Reblogged by slightlyoff@toot.cafe ("Alex Russell"):
mmasnick ("Mike Masnick ✅") wrote:

Attachments:

• He said that the platform had been on course to spend about $5bn in 2023. Overall costs at Twitter in 2021, the last annual period that the company reported before being taken private, were $5.6bn, during which time it made a pre-tax loss of $221,409. Musk predicted that Twitter’s net cash outflow, “if you didn’t make any changes”, would be about $6bn to $6.5bn next year. This is partly because the company has been loaded with $12.5bn of debt to help to fund his acquisition, requiring about $1.5bn a year in annual debt servicing payments amid rising interest rates, he said. “Not good since Twitter has $1bn in cash,” he said. “So that’s why I spent the last five weeks cutting costs like crazy.” His remarks suggested the company was on track to make about $3bn in annual revenues next year. That would suggest Twitter was on course for revenues as much as $2bn lower in 2023 than the $5bn it achieved in 2021 - which mainly came from advertising. Many marketers have pulled out of the platform since Musk’s takeover over moderation concerns.

slightlyoff@toot.cafe ("Alex Russell") wrote:

When you add complexity to the client side, you multiply costs relative to doing the same thing on the server. But none of the thinkfluencers are going to level with you about that.

At this excruciatingly late date it's still a pile of "could" and "can", rather than "does" and "do". Why? Because they know.

They know it's more expensive and isn't working, but their careers depend on you thinking they know the path forward.