msbroadf

I run virtualhere.com and i referred customers to the uhubctl list over the years so they can powercycle development boards and iLoks remotely that they use via virtualhere. I was tired of the recommended hubs changing internals (some models working some not) so i thought about creating one myself. So just went with a quality usb hub chip from Microchip that is actually documented (registers etc) and followed the spec and found a manufacturer in china to route/complete it. (Did the same with the other KVM/IP stuff i created).

The big issue i encountered is the hub programming software from Microchip is full of bugs and basically unusable and the docs are also a bit wrong. So i had to write a programmer from scratch to remotely program the hub chip registers so they are ready for manufacturing. E.g allocating unique serial number, setting up the chip fuses etc. That actually took most of the time. It programs everything from australia via virtualhere during manufacturing in china. I had trouble with the Great firewall blocking virtualhere half way through too! So i had to turn of ssl

The switchable host im not sure. I dont think many customers use it, I do have a few that use it for switching mouse and keyboard they plugged in. That functionality came with the hub chip so i just wired it up as it didnt cost extra. The hub chip also came with GPIO/uart routing but i didnt hook that part up.

I have a lot of astro users. But you need to use Ethernet for the connection between VirtualHere server and client and not wifi. A pi5 is very good for this.

Correct, i am a solo developer working on it since 2010 :)

Yes kernel knowledge and a lot of plain C. Surprisingly didnt have to read too much of the USB spec for the first few years.

Yes generally i reject any request if its customization just for one customer. I try to generalize the change by e.g adding an API call. Also i get requests to keep xp/win7 support which i reject often.

The biggest thing I learned is to basically do nothing on the product, unless its needed. I used to be tempted to change things/add features for fun and for a challenge but its a mistake. The reason is it might introduce bugs. Too many customers and big companies rely on virtualhere now day-to-day and the software is complex enough that its hard to tell if the bug was always there or just from the new addition. Im very relucatnt to even upgrade the compiler unless i have to for a new OS etc.

Fully Homomorphic Encryption

I had my large intestine entirely removed about 8 years ago, so no gut bacteria for me! And i haven't noticed any change

Yes its definitely viable. I started my micro ISV 6 years ago, its been quite successful about $200k/yr. Its very niche. I do no marketing but get regular sales through mainly word of mouth and internet searches. Now some bigger companies have picked up my software and use it in house and OEM it. Took a lot of work to get started and a lot of learning Its difficult for competitors to make something similar in a reasonable amount of time so its got a nice moat around it at the moment. I have no employees and almost no overheads so all revenue is profit. The most useful thing is when other companies build around your product, they provide valuable test / customer feedback which greatly reduces the work load.

No, it was removed for ulcerative colitis. If gut bacteria affected mood I would have noticed now I have a lot less but there wasn't any change.

I had my entire large intestine removed over several operations, and i felt no difference in my mood