I’ve been having a not-terrible experience with Osano so far

WeatherStar 4000+ Web based WeatherStar 4000 simulator that reports current and forecast weather conditions plus a few extras!

@chswx.com have you seen this? weatherstar.netbymatt.com

but see if the falcons had just made another field goal in regulation then they wouldn’t have had to go to OT, see? the league must not be questioned or its broadcasts reproduced without express permission

1. please don’t talk about my friend Claude like that
2. of course I don’t need it, it just takes me a long time to read obfuscated JS

so google can continue to receive conversion data even when your network connection is down, then retry the requests when you get access again. pretty sophisticated and elegant — if, again, not a little shady, even despite its privacy-respecting design

Service Worker API - Web APIs | MDN Service workers essentially act as proxy servers that sit between web applications, the browser, and the network (when available). They are intended, among other things, to enable the creation of effe...

service workers basically act as forward proxies that sit between clients and servers to intercept network requests and even work offline. they can't use XHR or web storage, but they can persist and run even when you're not on the page that registered them. developer.mozilla.org/en-US/docs/W...

the service worker uses HPKE to encrypt the conversion data and sends it via the Attribution Reporting API, and the 3p iframe provides an isolated execution context, so it seems like a good faith attempt to respect privacy. still seems ... I dunno, surreptitious in some way.

some of the service workers registered by Google Tag Manager

I discovered something about Google Tag Manager that surprised me a bit. GTM sets up 3p iframe service workers whenever it loads on a site, which act as proxy servers to Google data collection servers. This allows Google to continue to receive conversion data even when cookies are blocked

Pssst... Google, tell Ishita they forgot to remove debugging console.log() calls from the Google Ads Google Tag source code.

www.googletagmanager.com/gtag/js?id=A...

yesterday with Claude's help I wrote an algorithm to detect what type of Adobe Analytics track call triggered a doPlugins call and it's pretty slick

Since the analytics community is small here, let's spread the word and recommend people you can follow. Here's a starter pack/list go.bsky.app/Huxv35J

1 million users in the past 24 hours

waow

May my backups be current and my documentation clear
May I remember to test first, even when I'm sure
May I ask for help when I need it
May I share my learnings to help others avoid similar pain

(h/t Claude 🙏)

May my mistakes be teachers, not judges
May my systems be resilient, and my spirit even more so
May I approach problems with fresh eyes and an open mind
May I remember that every bug I create teaches me something valuable

agree on R framing of virtually any social issue at least at the national level, part of why I think I don't vote R. I'm just also realizing I'm against a lot of D framing of social issues. I (and others) don't think "if you're not with us, you're a racist/fascist/misogynist" is a winning message

I have this idea that the politicians I pick align with my values, but based on what exactly? I don't actually know their values! I don't know them personally. I don't know what they do or say in private. I don't know how they make decisions. I'm making specious inferences about how they will lead

I'm absolutely repulsed by most R rhetoric. I perceive it as reckless, churlish, maddening. I also agree with some traditional R policy inclinations, and I can't deny my R congresswoman has been doing a good job for our district and deserves another term. but what would it say about me to vote R?

it does indeed. but I doubt very many people have evaluated their candidates' proposed policies and made a choice based on which they believe would be better for the country. I doubt that I've voted straight D in every election since turning 18 because Ds have had better ideas every time.

such a great feature!

the good news is that now most people will have a more favorable view of our economy which is booming by most KPIs 😅 inflation back under 3%, unemployment at 4%, GDP up ~3%, wages up 6+%. ours is still the envy of the world

re: messaging, Ds are out of touch for most compared to Rs. they address demographic groups as if they were communities; Rs do not. most hispanic ppl don't like being called latinx; most ppl against biological boys competing in girl sports. assuming no inflation, not even clear Ds could have won

1. ppl are irrational 2. inflation kills incumbents (see recent incumbent defeats in UK, France, Japan...) people attribute economic success (or pain) to the executive even tho largely out of their control. 3. there was a huge crosstab shift R almost everywhere. most ppl not OK with D direction.

new blog post: been using Requestly for a few years now to debug Adobe Launch tagging, and it's the one tool I keep coming back to. just a huge fan of this product and the team behind it. ctysse.net/posts/master...

the sequence I recommend:
1. Consent Init: set defaults (either load from CMP cookie/JS var, or use static values from GTM)
2. Custom Event: update consent

for example, OneTrust sends a `OneTrustGroupsUpdated` event when a user updates consent preferences; that's what I'd use as a trigger for 2

part of why the Consent Initialization trigger exists is to allow you to populate some value if GTM loads before your consent management platform. consent mode doesn't *implement* your CMP, it just *integrates* signals from it with Google products

why not use Consent Initialization as the trigger for setting default consent?

I wish I had his slide making skills

worth it?

trying to convince a client in 2024 that they need a data layer makes me feel kind of insane

haha my exact same reaction. unironically. such a power move