Originally published at: https://tuxdigital.com/2020/07/thisweekinlinux-108/
Originally published at: https://tuxdigital.com/2020/07/thisweekinlinux-108/
FWIW, Linux Mint did minor updates through the Update Manager - this is how I upgraded from 19.2 to 19.3. I did watch a video the other night on “manually” upgrading from 19.3 to 20 https://www.youtube.com/watch?v=T4gYY_luRKo&t=306s but I will be waiting for the official method.
When talking about updates on Linux Mint I was referring to major updates only of 19.x to 20.x because every time they’ve made that difficult because of how they structured their distro.
19.2 to 19.3 is a simple upgrade because both are based on Ubuntu 18.04 so the core doesn’t move on those. Mint’s problem with upgrades is when it goes from core to core. 19.x has 18.04 core and 20.x has 20.04 core.
Linux Mint used to not be LTS locked so every 6 months the required fresh installs with no upgrades so they realized to stop doing that but since being LTS locked they still have the same exact problem, people just see it less often.
I’ve upgraded major Linux Mint versions to other major versions. There’s a command line tool they use called “mintupgrade”. A couple months back I tested this out because most people recommend doing a clean install for major versions. I went through 17.0 and it’s point releases to 19.3 without problems. Mintupgrade is just a package in the repo you download and run. Of course the package isn’t available until they have the transition working well, which takes time but there’s no huge rush, the 18.04 LTS is still supported until 2023. The Mintupgrade from 19.3 to 20 should be even simpler than previous versions because they don’t have to transition to a different display manager like with 19. It’s 3 lines in the terminal, which is NOT harder than upgrading Ubuntu.
It isn’t harder than upgrading Ubuntu from the command line but doesn’t Ubuntu give you the option to upgrade with a GUI? LTS users don’t get the upgrade notification until the first Piont release (which is around around the Summer), but then they only need to click a button. They don’t have to willingly look for it
When you say that the names that Vector chooses need to converge do you think that it could lead to confusion down the line?
Right now you can tell some one to use the internet and they realize that they can use Firefox or any other number of browsers to reach a website, using the underlying protocols. In a similar way, you can use a number of clients to use the matrix protocol. I feel like keeping the names separate would be a good thing because there are different clients (riot, fractal, ect.) and server implementations (synapse, dendryte, ruma, ect. ). I see where one could come at it from a perspective of lowering the barrier to entry though. Would it be better, as a community, to talk about particular clients then? It doesn’t really matter what home server implementation is being used or the fact that the underlying protocol is called matrix.
If the projects were called matrix-protocol, matrix-client, matrix-server and matrix-company it seems that it kind of takes away from future projects like the fractal client where everyone else will be playing second fiddle no matter what.
It’s copy and paste, whether Ubuntu or Mint you can do that with the mouse. There are some extra steps with Mint but the process doesn’t require any extra understanding besides “click this, click that” it’s at least as simple as using Facebook or typing an email.
I don’t think the point of the message was to claim that it’s “hard in Mint” but merely as a response you saying “is NOT harder than upgrading Ubuntu”. Copy and paste commands in a terminal vs click a button in a GUI. Yes, it is harder in Mint but that’s not to say it is ultimately difficult. The issue is that many people don’t know what terminals are at first and while sure the Terminal is nice to have and know . . . it’s not something that is as simple as clicking some buttons in a GUI. Honestly, if my dream came true, the terminal would be something no one ever had to use. I would still use it because of speed of action but if it weren’t ever required that would be ideal.
Nope because that’s been proven a successful path on many occasions.
I would argue that a lot of people don’t even know what a browser is. They just know the brand Firefox, Google Chrome, etc and that’s what they need to go to websites. Google Chrome is based on Chromium and a lot of other browsers are based on Chromium too but Google makes Chromium and Google Chrome has all of the power in that field. The alternative Chromium based browsers are just playing in Google’s sandbox.
These are very different things. The internet is decentralized but not really even by design, it was just made in pieces and therefore it inherited this structure. No one owns or controls the internet. It’s not a brand that you are trying to convince someone to use while already having competition at massive scale.
Matrix is trying to usurp existing players and those existing players are Slack, WhatsApp, Telegram, etc. This is a very big hill to climb so trying to climb it with confusion built-in to your own product is like trying to climb that hill while dragging 3 people up it behind you. (six feet apart of course, social distancing )
No, that wouldn’t be better and yes, it does matter which server is used. You can’t use these clients without having an account somewhere. You have to promote a place to have an account and a client to use that account. If these are competitors to the protocol itself then you also have a third thing to promote because they’ll want to know why this client and server is so much better than what they have.
You can’t really avoid talking about Matrix itself because Matrix protocol is the reason to even bother with talking about anything else involved in the equation.
Mastodon for example, people don’t promote individual instances or clients almost ever. Instead that vast majority of people just say “you can follow me on Mastodon” and have a link to their specific stuff when people search for it. Mastodon is the protocol but it’s also the brand that everyone uses and look at the size of the official Mastodon.social and compare that to any other instance. It matters.
Would these take away from the 3rd party clients and whatnot? Yup. This would 100% do that but the truth is that this already happens and they are already second fiddle. All of the clients that are alternatives to Riot are minimal features, platform dependent and so really the only client that anyone ever can promote is Riot. Riot is on all platforms including mobile and just telling people they need a Matrix account and the Riot app is what the vast majority already do.
The 3rd party clients are playing in the sandbox but they only have a limited amount of sand available to them. Is this ideal? I don’t know but it is what it is and because human behavior is what it is, that’s the “nature of the beast”.
Marketing is limited in what can be done and it has to have the right ingredients to be effective. Matrix has had way too many ingredients and each ingredient has to be explained to people before they can grasp why they should care. I mean I’ve been using Matrix for years and I am a very technical person but the amount of effort going through the maze of self-competition made me not care enough to really dive in. Noah dove in and found stuff that I had no idea about, this is a value but to reach mass adoption there should never be a maze to navigate.
Thanks again, @MichaelTunnell. As you’re offering more youtube content these days, I really should check there too. Learnt a lot from the Linux Mint / snap video
Cutie-Pi sounds interesting. Nice name, too.
On the todos front, I’ve been thinking for some time to code an inter-operating security-focused app for this, though I don’t find writing for Android and iOS all that appealing actually. I’d rather support UBPorts if possible and work on one that runs as native on Linux and on phones, though I think the phone side for this app would be the most useful.
yea you should
that sounds awesome! make the UBports app sync with Nextcloud Deck (or ToDos addon) and you’ll be my hero!
There are so-many apps that already exist any new one designed really does have to offer something that’s both new and desired. I’ll look carefully at the technologies you mentioned for some insights into their APIs. Even if initial build could work on Linux for starters I’m hoping getting it onto UBPorts would be the easier part of it. As you know so many apps, Michael, I’m sure the niche you identify would be a useful one to address and worth considering. I can’t promise but I can certainly investigate and I’ve been speculating on my own design requirements for an age, every time I use an Android app that doesn’t integrate with other Android apps and isn’t private due to reliance on some kind of cloud that isn’t self-hosted…
Awesome! btw I am not sure how valuable the nextcloud / deck solution is to be honest, because it is a very niche market. I think there would be a lot of value for a todo app or a notes app with sync. Another option would be to talk with the devs of Joplin to make a UBports version for that.
Thanks, @MichaelTunnell. Joplin looks very interesting indeed. I think with open source development a good strategy is to maximise reuse of existing work and minimise reinvention. If Joplin is already solving some of the problems we have in mind then a port of that could be a fine starting point. I’ll look into it
Michael, thanks for the full response.
It does seem that this is a bit of a different situation though. However, I suppose if there was something like matrix-company, matrix-chat, and matrix-protocol that wouldn’t necessarily prevent other clients from being used or promoted.
Right, they don’t need to know what a browser is or what ipv4 or ipv6 is. They can use Firefox or any other number of Chromium browsers. People only know the name of the client. That’s why I’m saying it doesn’t seem to matter what the name of the protocol or the company behind it is.
Maybe a better parallel to draw is e-mail though. If I ask someone to email me, they know that they can use gmail to communicate with someone on outlook. They know that doesn’t matter. In a similar way, I don’t think it’s that much of a mental hurdle for some one to realize they have options in matrix clients if I said let’s chat on matrix.
I can see though, where the e-mail parallel could be a little weak since times are different. When people learned how to use e-mail, that was the only public thing of it’s kind. Like you’re saying, now you have other chat platforms to compete with and they are really setting the precedence for how they should be used.
As an aside, I would like to hear why you think the internet is not decentralized by design. Correct me if I’m wrong but I’m pretty sure ARPANET was decentralized by design. There was no central server. As it became public, centralized services became the norm.
I see what you’re saying with Vector trying to compete with other companies. However, the Matrix protocol is just an enabler for clients. I can see too though that a single entity to compete with companies is better than a bunch of clients from a strictly naming perspective. I still think that there just needs to be a better way of talking about the protocol and the clients.
I was talking about the implementation of the server. That doesn’t matter as long as it is true to the spec. There should be more than one. That makes the spec more robust and better for everyone. It shouldn’t matter what home server you use either though. You can talk to people on other home servers. I feel like even if needing to put in extra characters to communicate with users on another server is a hurdle, that is a UI implementation detail. A client can make that easier. Your matrix username has the homeserver url in it. So if I was on a different one and put yours in to my client, it should be able to find you.
I agree about needing to talk about Matrix but I still don’t see why the names of everything else need to change. I still think that users can be led to use the clients on Matrix without a whole-sale renaming of everything.
With Mastodon I’ve seen the promotion of joinmastodon.org, which promotes other servers. The main mastodon.social is closed for new accounts and they link to either their new instance or the joinmastodon. I agree though, there should be a main one to promote and that a server with matrix in the name would help adoption. The nice thing about joinmastodon though is that it is more obvious about how to join than the Matrix site.
Yeah there is no other client that is cross platform but I can’t imagine that always being the case.
I agree with you that it is hard to get started but really I think that could be solved with a better landing page at matrix.org or somewhere else like joinmatrix.domain or enterthematrix.domain. The main matrix website could stay as it is and really promote the protocol. That could promote a client and home server and say, look it talks over something called matrix, it’s like email, click this button and get started. All this to say, I don’t think that someone even needs to be told what everything is up front, that’s what the main matrix site is for. I get you’re a marketer and have expertise there but maybe just presenting it better would be sufficient.
I guess I see the points your making but still disagree that everything needs to be renamed.
One interesting thing though, could be a study. It would be interesting if a couple scenarios were examined.
- Fork everything and rename it for these purposes.
- Have a fake company website about the protocol spec and repos for the client and server.
- This would have the benefit that if people learned about everything this way they would just conflate the protocol, server, and client as just Matrix so if someone says “lets chat on Matrix” they might not “know” what they mean but they could guess and just use the matrix-chat application.
- People aren’t even told what matrix is, just “let’s use riot”.
- This could be interesting because it would be kind of a mix of both of our positions.
- The Matrix protocol isn’t even talked about really on their front page. Just the client, so if you only know riot it is kind of like Telegram and Slack where you only know about the application.
- This would have the downside that people here would probably be confused if anyone talked about the protocol and said “let’s chat on matrix”.
- There is a better landing page like joinmatrix.domain or matrixchat.domain
- This would be kind of what I’m advocating. Just talking about it differently but leaving names intact.
- A main Matrix homeserver (where the chats happens) is promoted and a main Matrix client is promoted (the way you chat). Maybe other homeservers and clients are promoted on a second “explore” page or something like joinmatstodon.domain.
- Just a dead simple page without much other information, allow main matrix site to retain the gory details.
- The current situation
- Which is confusing. If someone says you should get on Matrix, I feel like most people would go to the main matrix site and we can all agree that is dense and confusing for a technical and non-technical people.
For each it would be interesting to see if people could get friends and family to use matrix and which situations where the easiest for people to understand. Obviously this would be a lot of effort and not realistic as it would need a lot of participants and the fake infrastructure alone would be a nightmare. None the less, it would be interesting.
Anyways, thanks for the replies. I see where you’re coming from.
[Edit] Also, I do enjoy your show and would like to thank you for all that you do for the Linux community. I just heard this on your show and just wanted to discuss it a little.