I generally use Ubuntu, and as I’ve been having issues getting Mycroft to work on Ubuntu was wondering if there was a ppa to make it a cinch to install. I was initially pleased to see that there is indeed a ppa, but very dismayed to see that it hasn’t been updated in over 70 weeks. The kind of issue that I’m experiencing isn’t one that is insuperable, but rather of the nature of being a proverbial pain in the posterior, and one that would be enough to drive someone to use an alternative, one of the competition that is effectively fit and forget - namely the Google, Amazon or Microsoft variety. This basically means that Mycroft, for all its very laudable qualities, such as respecting privacy and being open source, is a fail in that it’s not fit and forget. True, using the alternatives means that I’d have to buy yet another box to have somewhere on a shelf in my living room, or wherever else I use a voice assistant, and which has been suggested to me as the easy work around for Mycroft too. The weakness there is that Mycroft is still very much a Work In Porgress, and things are likely to change quite radically as the system develops - and besides, I’m someone who is very much trying to go the convergence route so that I avoid a proliferation of boxes in my living space. I want to run the system on my one main computer that I use for a multiplicity of functions - everyday desktop pc, media centre, gaming machine, graphical design workspace and a host of other functions. I’d also like to use this machine to run a voice assistant. If needs be, I will mess about until I get it to work, but realistically, unless Mycroft is pretty much install and forget, it will be something that is very much on the back burner for many potential users.
I understand that it’s a project in flux, as it develops, and I am contributing towards the translation, as it is the only voice assistant project that is likely to be available in my language. However, if there isn’t a non-techie easy-to-intall version, many would be users, who could provide valuable user feedback to the project, just through their everyday usage, will go elsewhere. There is of course a vital role for those who want to be more hands on, and techie in their contributions, but I think the point I make is one that should be seriously considered. On balance, the very fact that Mycroft is unique amongst voice assistant projects to respect user privacy as a fundamental aspect is probably a USP that should swing it as far as many are concerned, but unless it’s easy to use, by those of us who want it on our desktops, as well as in in yet another box, I suspect that people will downplay the privacy aspects and choose the commercial alternatives that spy on every aspect of their lives.
I like your point, and luckly it is posible to get a mark_1 and soon a mark_2 that works for the endusers hwo just need a smartspeaker whitout sharing privacy at the same time.
As far as I know mycroft hassnt a ppa for mycroft, so if you have found one, it isnt mycroft who are maintainer of it. Or correct me if I am wrong @KathyReid
That beeing said, it would be a good way to make mycroft mainstream with repos for the common Linux distroes. But I think mycroft should concentrate on getting the software even greater, and others could maintain repos.
Main issue with mycroft seems requerements and most of all setting up sound and mic.
My personal view of this project is it is very much still early beta stages. There is a great community of enthusiasts with very different skill sets and ideas working together to make this project something that people will want to embrace. There are many great skills already and many more great ideas that will transpire as we traverse the road-map. I am not sure we will ever be able to (or even want to) keep up with the “big data players” developing products with the sole intent of scraping growing amounts of your personal “metadata”. I am confident that we will reach the point in the not too distant future where mycroft.ai will be a product that will be capable of leveraging the growing knowledge of the community to interact with its users, its skills, and your devices to create a helpful user assistant who’s primary function is to assist you, as opposed to a trojan horse who is more interested in collecting your personal habits than making mundane tasks easier.
It’s still early for everyone. On a typical day of scrolling through the abstracts at arxiv-sanity.com, there are at least four or five papers reporting that they have advanced the state of the art in machine learning in a significant way.
Thanks for your feedback @MZTS250Supa5.
This is one of our key dilemmas; of course we want to make Mycroft as easy as possible to install for as many different types of distributions as possible. However, we’re a small team, and can’t take on the maintenance overhead of maintaining packages and repos for a lot of distributions.
That is why the
dev_setup.sh file attempts to install dependencies for multiple different distributions - it’s a “best effort”. Even then though there are several distributions where this doesn’t work - for example, FreeBSD.
I’ve seen many forum posts like this - I wish Mycroft was easier to install, when is Mycroft coming for $MY_FAVE_DISTRO, can Mycroft please have $FEATURE. Unlike Google or Apple or Microsoft or Amazon, Mycroft is a very small company, with a tiny staff. We cannot make Mycroft work perfectly on every distro, or package for every platform. What we can do is partner with the Community - you - to provide the tools for others to contribute to packaging and installation efforts. Our open source Community is our greatest asset.
So, how can you help improve Mycroft?
There are several ways. You can contribute to mycroft-core, if you know Python you can write Skills for Mycroft, or if you have a half an hour, you can help train our Precise Wake Word listener or help train DeepSpeech.