eap_proxy VS Dash-Board-for-Newton-OS

The inspiration is the EAP Proxy that I used to bypass the AT&T router and connect my router to the AT&T fiber, before I learned about the vulnerability that allowed me to copy the 802.1x certificate from the AT&T router to my router. https://github.com/jaysoffian/eap_proxy

I use an EAP proxy to put my router in front of theirs and got a %20 speed improvement.

With my wife, I built two kids both in college now. That's definitely the most impactful.

In the online world, I've contributed bits and pieces to open-source here and there as far back as the late 90s. I think my first contribution was to the shadow package, but I've contributed here and there to Apache, Radius, random packages that I use that I discover bugs in, etc.

A python script I wrote to allow folks to bypass AT&T's residential gateway was used by more people than I ever expected:

https://github.com/jaysoffian/eap_proxy

There are forum threads about people rooting these routers to try and extract the certificate/private keys from their provided router, which worked for a limited time with certain router models. For now, the most popular way to do it is with a tool like this: https://github.com/jaysoffian/eap_proxy

If you use a Ubiquiti router, then another method to bypass your ATT router is called EAP Proxy.

WEEELLLLLL, there is a way to get into a true bridge mode on some ATT fiber plans/service by using something like EAP-Proxy: https://github.com/jaysoffian/eap_proxy

If you're up for a more complex setup, it is possible to put your home network on AT&T Fiber without going through their boarder gateway. I've had luck using https://github.com/jaysoffian/eap_proxy to bypass the boarder gateway for traffic flows.

For me, it was the second application I ever released, when I was a student at university and still didn't really know how to program properly.

The application was Dash Board[1] for Newton OS, and it only ran on the final generation of Newton hardware (created by Apple, but spun out as a separate company in its final days, before being killed by Steve Jobs shortly after his return).

It "only" sold a few thousand copies. (But it was during the warez heyday, and I am pretty sure there were also tens of thousands of bootleg copies being used, thanks to the registration code generator by "DocNZ" that was widely shared on Hotline back then.)

But that was really pretty great, since the final MP2000/2100 generation of hardware it required was thought to have only sold about 200,000 devices in total.

I have since had a fairly normal software engineer career, and have worked on apps that shipped far more copies, and today I work on customer facing web applications and API SDKs that have more users, and arguably do stuff that is more "important" (e.g. help companies manage large fleets of machines/robots/IoT stuff) than what Dash Board did — which was basically just improve the user interface of the Newton.

But it's 100% clear to me that the magnitude of user impact of Dash Board was much higher than any other thing I've built. People really loved it — I know because hundreds of them actually wrote to us to let us know. (LOL I mean wrote to me "me" — old habits of pretending the company wasn't just one student in his tiny apartment die hard).

Of course, I made more money later, and worked on things that touched a much larger number of people's lives. But "impact" has both X and Y axes. It was the depth of the users' fondness for Dash Board that makes it eclipse everything since. I don't think there are that many chances to just go for "user delight" as the number one metric.

For me, developer satisfaction is a function of that user delight more than anything else.

[1]: http://www.fivespeedsoftware.com/dashboard

[2]: 15 years later, I open-sourced the code and gave it a proper retrospective: https://github.com/masonmark/Dash-Board-for-Newton-OS

The experience of resurrecting this software is similar to the odyssey of Mason Mark, detailed here.