Mark II Update - March 2020

I’ve been listening about the “dead of the desktop” since more than 20 years now. How consoles would kill the desktop, how mobile devices would kill it… and here we are, both of us writing this from a desktop computer. :wink:

Oh sure, I know the argument that everyone has a mobile phone, but not everyone has a desktop.

A desktop is not necessary to browse facebook|instagram|put your un-social network here, or take photos of yourself with a duck-lips pose, but being sincere, there is not yet a real alternative for a desktop. At least in the foreseeing future. I cannot imagine all the enterprises working with excels in a mobile device, developing with a phone, or working in an Intranet with a tablet. These devices are completely useless on the enterprise environment. And in our homes, before the desktop death, I see a new desktop Renaissance. Perhaps with newer human interfaces, like voice assistants or augmented reality, can boost desktop experience. Perhaps we could stop calling “desktop” but “computer”, “home”, or whatever embedded computer in your home. Some kind of computer with a screen you can interact with.

And even if the desktop were slowly dying, years of which have I been listening this made me skeptical, what is the problem in a desktop installer? I mean, as broader is the offering, broader will be the audience, am I wrong?

1 Like

If only MaruOS would become mainstream, you both would be right :wink::joy:

https://maruos.com/

Its a device and that is what makes me wonder about snaps and maybe Ubuntu has this wrong.
Devices are partitioning out applications to provide certain and specific roles and I am wondering if the whole packaging of applications we knew will be more like Mycroft is now where its headless and can use an array of GUI’s.

The question is there any point in a spending the time on a specific snaps installer for a specific desktop.
Some think prob not as Mycroft is a voice activated AI driven HMI that should be positioned away from any specific desktop implementation.

You just need to google usuage stats and how people commonly interact with technology.
The desktop is dying, if slowly and will never die.
https://www.broadbandsearch.net/blog/mobile-desktop-internet-usage-statistics

Its likely if we are going to blur the word desktop then a home AI is going to interact with many device desktops and that a home AI might even be a centralised home server.
Hence why I keep name dropping the likes of HASS.IO, OpenHab & Domiticz as the co-ordination of a home network of devices might well be where the money is of the future, action or just pure geekery.

There are so many competing technologies so many possible desktops that a choice of any singular is statistically likely to fail.

I would stick with the headless and make it available in containers but after reading also wondering if this should be via snap (headless).

Sonos snapped up Snips :slight_smile: and its interesting to see where they go.
I think rather than try to pick one from many there is just a few homeservers that speak, display and connect to many.
So by focusing on a few you enable many and once more name drop hass.io, openhab & domoticz

Well, it’s on the Ubuntu Touch roadmap https://ubuntu-touch.io/features/convergence :stuck_out_tongue:
And PinePhone in fact runs desktop applications, when the HDMI output will be working we will be able to use Purism’s desktop application convergence :smiley:

Aha! That’s the whole point.

I would also rather prefer mycroft guys has invested their time in other efforts, like @j1nx pointed in his post, not to create a snap file which is heavily targeted to ubuntu. That’s why I upvoted for AppImage, because at least AppImage has everything bundled on it, and since they’ve done it, it would at least have made more sense to provide an installer for any linux distro with any desktop (or without desktop).

On the other side, Mycroft has always had in mind the possibility to be used in combination with visual aids. Mark I had a plate and Mark II has a screen. If it has become popular amongst rpi user only shows the potential it has both with and without screens. Is the default voice assistant in Plasma Desktop and Plasma Big Screen, and it has several skills only useful if they are ran on a desktop.

Sorry, but again, when we could do whatever we do with computers besides browsing the Internet, we can speak about the death of the desktop. Smartphones devices are created just with this idea in mind, and desktops are still quite predominant after 12 years.

Well, snap is more or less a container. Is not as popular as docker and I don’t know if has the same security benefits, but if you’re an ubuntu guy, you will feel more confortable having all the stuff “snapped”.

I will continue using Mycroft in all its possibilities: on Mark I (and Mark II when the day comes), on all my desktop computers (all with arch + plasma), my magic mirror (need to find a proper mirror and assemble it), and when Plasma Big Screen improves a little more, to get rid of Amazon FireStick (because my 1st generation firestick, since the Alexa “upgrade” it began to fail and become slower). If someday pinetime see the light, I will be glad to use a mycroft on my wrist as well.

That’s why I’m glad to see any advance towards any field. I’m with you I had been pretty more happy if they had released for example multi-language support natively, or shown any progress on the Mark II device which I backed and I’m patiently waiting for, but hey, a desktop installer is also welcome.

1 Like

That blog post from a RedHat employee, a bit old now but anyway nothing has changed since then, speaking about the snap/flatpak universal installers.
https://www.happyassassin.net/2016/06/16/on-snappy-and-flatpak-business-as-usual-in-the-canonical-propaganda-department/

Interesting reading

As I got tagged, please let me explain;

I am not dissapointed that Mycroft is focussing on a packaging system. I have no oppinion about that and will work with whatever they decide.

I am dissapointed that apparently resources got stretched so thin that software and hardware trajects are not running side by side anymore, because in my oppinion the packaging of the end software results has nothing to do with the hardware. At least not at this point in time, because the decision about which packaging software to use, does not influence the decision of the hardware. It actually worries me a bit.

We haven’t heard anything about the hardware traject anymore since January and we are still awaiting the big word about the partner and where we are currently at.

Although dissapointed and a bit worried, I also know EXACTLY how those things go. I had my fair bit of hard business decisions myself when I put “The Little Black Box” onto the market. Many, many moons ago.

It is actually the reason I lost most of my hair :smiley:

Edit; I have to add. Being a bit disappointed doesn’t mean I blame them for something. You can only spent money once…

3 Likes

I didn’t expect to put words in your mouth you never spoke :wink:

I’m not upset at all either, just, as you if I understood for your first post, I expected some Mark II’s update, and to be honest, talking about a package manager seems to me unrelated to Mark II’s roadmap.

If they progressed on the roadmap, hasn’t been shown on the entry. It would be great some information about the status of the new device in every Mark II update, I’m happy with a single line telling us if is stuck, and we need to wait longer, or if the process is advancing slowly, just to not leave us backers in the dark.

As I said, I’m not upset nor disappointed, and I know dev team and community members are progressing steadily, making software even better each day. Just expected a Mark II’s update in a Mark II update.

3 Likes

If you read the posts about patent trolls there is prob a very good reason for the delays in the Mark II.

If I go back to the OP and don’t get distracted then there is no importance in snaps apart from Canonicals release process for updating devices that uses snaps.
So yeah in terms of devices that is hugely important and apols as in a forum you post on the last post and quickly forget the OP.

There is also https://www.balena.io/ most will know etcher but if you have gone snaps then fair enough but they are very raspbian orientated and also do container based fleet device services.

Consumer grade fleet OTA and the services needed are very dependent on resources so if you think the fleet services from Canonical provide what you need then fair enough from me as its a situation I can not envisage.
I am thinking for value services that do the same with a raspbian base https://www.balena.io/ might be an option.
You might have a voice and an element of leverage with them, due to both being smaller opensource with strong parallels.

So prob from the forum there isnt really consensus or preference to container as the rationale for snaps really isn’t about the container but purely a question of is that a suitable package / system for consumer grade OTA?

Then dunno if anyone has read my adventures of cheap hardware through a s-pipe.

No-one is talking openingly about hardware just a few have gone to huge effort to say ‘I am doing this.’
There is a massive oversight to something openhardware has been missing which is crowd prototyping but the focus on a singular finished complex product in the manner of operation seems a complete disconnect to the great manner of software development.
Tried to say it many times now about modular kit forms, using personal specifics as an example.
That R+D isn’t just this huge cost its something you could sell.
Really not sure why you are kickstarting so high, aiming for the last hurdle before the first has been straddled.

I am going to stop there as my usual bemused and confused for once is a massive understatement.

i guess this relates to the mark2, like the post says you cant expect users from a commercial product to “clean git orphan objects”, but for me this raise questions, so its the mark2 running ubuntu? does this mean the mark2 will install its software by snaps?

isnt this a bit limiting, since end users dont really care how this happens, why not go with docker which runs in more places/Operating systems? if you are devoting time into this then make it in a way that benefits more users… a snaps container is useful for ubuntu users and i assume mark2 users, a docker container would benefit everyone.

I’m curious to hear the rationale on why snaps is a good investment of developer resources, what is the benefit you are after exactly?

2 Likes

From the perspective of a desktop user on Fedora, using KDE, I prefer everything to be either the system package manager (in this case .rpm files) or to be flatpak. Thankfully, snap isn’t hard to install, but I find that they force auto updates, and I prefer to have control over that.

My experience with mycroft over the last year has been greatly hindered by a lack of update tool that doesn’t feel like I need to read through a huge shell file that touches files and folders all over the place. I personally prefer to use tools that make updating all of the software on my computer very easy, and I would definitely install mycroft in at least 3 more places if there was any sort of package management, aside from docker. To be frank, docker’s nice, but it’s not something that end users would know what is. An outsider’s perspective of how docker works seems like commands with less than 20 characters to install, update, etc. are few and far between, but that’s just because I’ve never seen end user software shipped in a docker form, so we need to keep in mind the intended use case for each platform when making a choice.

From what I can tell, these are the targets of each solution:

  • Docker is, from what I can tell targeted for web servers.
  • Snap is very dedicated towards Iot, as the ubuntu newsletter says
  • Flatpak is targeted for desktops
3 Likes

@J_Montgomery_Mycroft
Will there be a Mark II Update - April 2020 soon…? I did not hear a lot about the real Mark-2 device progress in the last time :frowning:

4 Likes