An interview with the creator of the Tauon PC-1, Volodymyr Mishin
I'll bet you don't roll out of bed and think about making your own PC before breakfast. Even you do, I bet you haven't actually shipped. Here's someone who has.
I reached out to the creator of the Tauon PC-1, Volodymyr Mishyn who kindly took more than a little time to answer some detailed questions about the PC-1, how it came to be, the challenges of a project like this, his future plans for Tauon OS and what is clearly a very personal labour of love.
Many thanks Volodymyr.
Your news blog suggests 30 days to prototype, and 6 months to ship! If that is right, it is very impressive, given the technical and logistical activities that had to occur. Is there more to this story?
I was looking for available options before I started the project. I got myself a couple PI (3B, Zero) OrangePi, C-SKY. And tried to build some kind of educational PC for my kid.
Unfortunately, the performance was awful, using Linux was too complicated even for me (I have no experience in desktop Linux. However I have some in server administration, can setup LAMP, nginx etc.)
Also, this was not my first attempt to launch something in production. I had previous (failed, in all regards except for experience) project - RevolVR. The biggest delay was with case molds, it took really more time than it should. So while most of the project was done in 5-6 months, just to make casing took additional 3-4 months. And I don't consider it perfect.
When did you start shipping units? How many units have you shipped?
The first computers were sent at the end of January 2020. Some of them were sent to YouTubers, some were sent as samples to governmental institutions of small countries. Some were sent to my alma mater school.
In total, for today about 500 units were distributed. About half of them for free.
For you personally, what would be a nice number of shipped units for the project to hit?
I have 2000 pcs already produced. If I manage to distribute all of them (not necessarily for money) I would be happy.
How big do you estimate the user community is?
Not big enough yet :)
Are you where you would like to be at this stage of the PC-1’s life?
All that coronavirus situation of course affected sales a lot. I had a couple of interested parties to whom I might sell 1000+ PCs. But now all these developments are suspended. Also, sales on the USA market dropped almost to 0. Before this "pandemic" I was getting 50% orders from USA and 50% from EU/UK. Now its 10%/90% . This is in terms of sales.
However, other my goals were achieved: design, manufacture and get the real product to the market, impart interest to computers to my son. So even if I will have to distribute the rest of the computers for free as charity - I will still consider this as a successful project for me.
Do you have any data on your users, eg: age, where they are from, why they bought a Tauon PC-1? Has the data matched your expectations for who would buy the PC-1?
No, I don't have these kinds of stats. About half of the units sold as "Educational computers", with another half being sold as "Computers for retro-gaming". Hopefully, at least some of them will teach some kids on computing.
There are only 5 people in your team. Alexander must be at school mostly (or maybe remote learning), and Vladimir must find it hard to code in the dark, so there’s really only 3.5 of you. Is this a full time job for all of you?
Big part of the work on the project is done by me. Including case 3d model, and even manuals design. I hired people only to resolve specific tasks for a very limited time. In other case, I could not offer this product at such a low price. I'm working 14/7 on this project.
It must be difficult with such a small team. If you could have more help, what areas would this be in?
At this point, I don't see how bigger team could help me to move faster. However, more testers giving feedback would be highly appreciated.
Have your Chinese partners been able to provide you with all the technical information and support you need, to develop for this board? Is there any technical information they are unable to provide, for legal, commercial or technical reasons? Has this held you back in any way?
They really helped me a lot. However, sometimes it was faster to do reverse-engineering than get help from them. Not because they didn't try to help, but because they don't have deep knowledge of the board and firmware.
Allwinner is selling the boards and provide an SDK. Sunvell usually just makes cosmetic changes for the final customer. Also, they are very busy with big orders, it's basically a miracle that they decided to spend their time to help me at all, considering such a small quantity. I am highly appreciating this.
All Chinese companies I'm working with are great, all tried to help a lot. And none tried to scam or fraud me. This is absolutely amazing - if you tried to run a "real world business" you would know what I mean.
As so much depends on your personal efforts, the users will want to know: how sustainable and resilient is the company?
At this point, unfortunately, the project completely depends on me. Hopefully in the future with the community growing bigger this would change.
The PC-1 is meant to be a simple, affordable, family computer. Yet, modern PCs and phones are already affordable, and can do everything the PC1 can. If you already have a phone or PC, you don’t really need a PC1. If you have neither, you would be inclined to get a “real” device first, instead of a PC1. So, while you consider this a general computing device, has the truth of this difficult logic meant you are selling more to a niche, not a general user?
Of course, this is niche product. For geek parents and geek kids. I personally don't consider the skills of using the touchscreen as something worthy. And you don't have a lot of "real" computers in this price range.
Why not cater more and market specifically to the natural users of this, the hackers, makers and amateur coders? You already encourage people to put different boards in the case😊. And the Allwinner-H3 is an OrangePi, after all.
I'm trying to reach all audiences. Hopefully, the more the community will grow, the more people will know about this product.
You have clearly put a lot of work into TauonOS. Do you have a roadmap for the OS? What is your ultimate vision for it?
The ultimate version of it will contain much less of Android itself. With the ability to launch both native Linux and Android apps.
Because it is a customised Android 7.1, at some people who don't look carefully will see this as a “lazy” effort to create a “quick and dirty” computer, with minimal development effort. This must be upsetting. Can you shed some light on how much technical effort was involved in getting TauonOS to where it is today, and how much remains to be done? What are the challenges in building something like Tauon OS?
The main problem is the huge size of the codebase. Fixing or changing anything in 25+ Gb of sources is not an easy task.
A big part of the effort was put into optimizing AOSP [Android Open Source Project] and "fixing" some (intentional?) bugs in AOSP.
After working on this project, I think here is Google's monetization model: yes, you can use AOSP for free in your project. But it has so many little problems and slowdowns, that it makes your product unstable and almost unusable. Want all this fixed? Pay us money and we will provide patches to fix all these issues.
That is why many Chinese products based on AOSP are so unstable and buggy. Lack of documentation on internal AOSP structure and confusing (intentionally?) comments don't make the task of fixing easier. Some parts of the AOSP seems to be written by undergraduate school students. Making the mouse work properly and adding "Gamepad" mode wasn't easy at all.
Are there less obvious features of Tauon OS that you would like users to know about?
Apart from "gamepad mode" in the future, I'm planning to change the behavior of the mouse, so that left click and mouse down would be considered by the system as touch input, making more software compatible with Tauon PC.
Due to the horrible architecture of the input subsystem, this is not as easy as it seems. I wouldn't promise this anytime soon, as this task has low priority to me. The next plans for the OS is to get rid of SystemUI monster process, which consumes a lot of resources and has no use to me except of task switching and task-killing.
Why did you choose to customise Android, instead of a mainline Linux distro?
I don't consider Linux in its current state as OS for beginners. As I stated earlier, I tried to use Raspbian for the purpose of teaching, but it's impossible. Even to install or find software in need - its a mess. Android on the other hand has a lot of educational software and games built for it. Also, the manufacturer of the boards didn't have Linux distro for it, nor had tools or experience to help me. If I decided to base this project on Linux, it would probably never finished.
This is a pretty unique product for the simple fact that it comes with a prerooted Android 7.1, and has a keyboard. Can you point to a toolchain for Android app development that would run directly on the Tauon, so this could be used as an Android development device, running on Android itself?
At the moment you can use 2 programming languages - Mobile Basic and N-Ide Pascal to write code on Tauon. Also, I'm working on Tauon Basic - a very simple framework and IDE, based on JS. While it doesn't allow yet to compile code into binary - this feature will be available in the future.
The good thing in this approach is that with very small changes, the same apps will run on Linux, Windows, Android and Mac OS. You can check an alpha of the language (in your browser) here:
examples you can find here:
Can you point the community to any resources for beginning Android development on the PC-1?
Any apps made for Android 7 should run ok on TauonPC, if they can be controlled with mouse, keyboard or gamepad.
There is already a mainline Fedora32 Linux image which runs well enough, but is still “unofficial”. Linux is at the front of many of your user’s minds on the forum for sure. Do you have any plans to officially support a common Linux distro?
The biggest drawback of the current Linux distro is the lack of drivers for NAND storage. I don't consider it as a usable alternative without this.
Are you familiar with the Maximite computers? Or, ironically, the “Tau” bare metal OS? This datasheet: documents so much low level information about the hardware, that I can’t help but think that it would be possible – not trivial - to boot straight into a “modern” 8- bit “OS” experience, like a Spectrum or a C64. That is, turn it on and boot straight to a blinking cursor, a programming language as the “OS”, and with the ability to make direct calls to hardware, or at least to the storage, sound, graphics hardware, and at least one external interface. Any plans for anything like this?
Yes, I'm familiar. But I think this is even more geekier product than mine! As for bare metal OS - I really have enough work to do with current OS and TBS.
Also soon some funny firmware would be available, based on TauonOS. I have sent a couple of devices to the developer of ZXBaremulator for RPI. However, he warned me in advance that it would be impossible to make ZXBaremulator compatible with my hardware, so I wouldn't expect any bare OS any time soon.
Can you please get Termux to work? I’ve had no luck😊.
At this moment I really don't have time to spend on this. You can reach hentacler on the forum in this regard. This guy has written the whole Wayland manager for android, "Sparkle". he probably could help you with this.
Why not make Tauon OS open source?
I can give access to sources by request. However, to assemble OS I use another tool, provided by the board manufacturer. They asked me not to share this tool, and I will not do so. So while you can compile and assemble Tauon OS from the sources, it would not compile into the identical firmware.
Where would you like Tauon to be by the end of this year?
I'll continue the development of the OS and tools as long as I am able to do so. The bigger question is where will the World be at the end of the year. I'm afraid we haven't even seen the beginning of the crisis yet!
Is there anything else you would like to share?
Hope I answered all questions but if you have more - send them right away.
Sorry for my poor English, I don't feel to fix all mistakes in that big text) Thought it would be much shorter.
[Volodymyr, your English is perfect, and it is better you spend your time rewriting AOSP:) Thank you very much again!].
It's great to you posting again.
ReplyDeleteI've been following this blog since I bought a Dingoo a380 several years back. I would be great to see your take on the newer handheld emulators, and other single board computers, like the Tauon PC-1 and others being released lately.
Hi Eduardo, thank you, that's nice of you to say. I'm glad there are still some people out there keeping the flame burning for these devices! Keep well.
ReplyDeleteThis comment has been removed by a blog administrator.
ReplyDeletePros • Accurate pH levels and urine contents • Comes with own heating strip • Very affordable price • Relatively long shelf life Cons • Not meant for beginners • Doesn’t come premixed Aside from being a great affordable option if you’re looking to pass that drug test on a budget, Test Clear’s powdered synthetic urine can go toe to toe with some of the other top brands in the industry. The fake piss contains what you’d expect from natural urine: accurate pH levels and the right consistency, natural compounds like urea and uric acid, and the heating strip ensures you get the right temp when you’re ready to hose down that cup during your test. And although it’s not premixed, which means newbies might find the process jarring, Test Clear provides clear instructions with their powdered pee kit to make sure that you use it correctly. Even better, its powdered composition means it has a longer shelf life compared to premixed urine, so you can stock this relatively longer for future use. Pros • Accurate pH levels and consistency • Comes premixed • Possesses natural pee chemicals • Includes an extra fake pee pack Cons • No applicator included • Short shelf life • Heat source not included Being high as balls means that your reflexes might not be at optimal performance condition, and accidents could happen if you need to handle some synthetic pee with you during those moments (or any other times for that matter). Quick Fix 6. Visit: https://www.urineworld.com/
ReplyDeleteReally this is very nice article. thanks for shearing.
ReplyDeleteDesktop or Computer on Rent in Delhi