Thunderbird.net Has a New Look | Hacker News Reader

Thunderbird.net Has a New Look

blog.thunderbird.net

Been using Thunderbird for a few years after using mostly webmail and it has been a mostly good experience. The UI revamps have definitely split the room somewhat, but honestly, Thunderbird's UI progression feels much better than Firefox's, and I have actually gotten used to the way Thunderbird looks and functions without too much pain. The information density is still generally alright, the tabs look like tabs, etc. The only thing that really tripped me up is the inconsistent hierarchy between the tabs and the sidebar, and I am not really that torn up about it, to be honest.

What I really think Thunderbird could use is account sync. For me on NixOS (how do you know if somebody uses NixOS? Don't worry, they'll tell you) I might just set up some secret data in SOPS and have an activation script ensure the accounts are all present, and then IMAP can take care of the rest. But for the general user, and maybe even still me, Firefox Sync support for accounts and extensions and more would be amazing.

Speaking of IMAP, JMAP support is another thing I'd love to see.

I believe account sync is among their higher priorities: https://blog.thunderbird.net/tag/thunderbird-sync/

I am aware, since I was looking forward to it in 115. Unfortunately, I kind of suspect it might not really be on the way, as last time I looked into it it seemed like it was in a bad state. I guess it's not a big deal, though.

"on NixOS" is the new "arch btw"

There is one difference. Arch users will tell you, "I use Arch. You should try it." and NixOS users will tell you, "I use NixOS. You should probably not try it. You should try Arch."

For me it's more often than not been the opposite. Although it's not just that you should use nix, you need to use their specific subflavor of nix features and anything else is heresy. It doesn't matter if the way it's documented is a mess, it makes sense to them.

I'm pretty sure that picking a religious denomination is less aggressive than hearing nixos users talk about how superior their system is.

Well, you won't hear this from me. Right now, NixOS is probably not worth the time for the vast majority of people. It's hard to quantify because while the benefits of Nix are immensely clear once you have things working, the amount of effort it takes and the complexity you have to deal with to get there is downright insane. It's not that I think other people are simply incapable of understanding it, not at all; I think people are but I'm not sure if everyone who does will be pleased with how they spent their time.

However... There are people among us who don't care about whether something is "worth it" or not. We want a complete, 100% solution to some of our problems. Not a half-measure, not a hack over the top, a from scratch, 100%, complete solution. What NixOS and GuixSD do solve, they solve very, very completely. The problem is that doing it is very hard, and in my experience Nix sometimes makes things much harder and more complex than it necessarily needs to be. Of course, making things as simple as they possibly can be is hard work, and so I don't blame anyone.

A good example of a system where people are clearly unhappy is kernels, the way kernels are built in NixOS is pretty unsatisfying, and the knock-on effects of it can be pretty confusing to new users who wonder why changing which kernel they use makes it so that none of their kernel options are applied anymore. (Silently! No error, just, your kconfig options stop working!)

This is on top of the fact that the community is split on things like flakes, leaving Nix in a bit of a precarious position, where there are multiple solutions to a problem and yes, people don't agree on which one.

And also, it's built on top of functional programming paradigms, which I contend makes a lot of sense when everything clicks, but that's an additional hurdle for the vast majority of people, even programmers.

Even if Nix was as simple as it could be, what it tries to solve is a very wide problem space that is complex by its nature, and it is harder to hide that complexity the way that many Linux systems try to. So, I don't suspect it will be worth it for most people to bother learning Nix. I think that the people who know they want Nix are probably going to figure out it's what they want eventually. There's a type of person who will invest in solar panels happily without caring much about the break-even; I feel like Nix is a bit like that.

So why do people evangelize NixOS so much, when surely they know it's vastly complicated? I can't say for sure, but I can tell you one thing: when NixOS first started to "click" for me, it felt like I was experiencing a future of operating system design that is still years off. Whether Nix is the future, remains to be seen.

What sold NixOS for me were setting up servers, and I would also recommend people to set up their *servers* with.

It was when I watched someone transfer their server to another disk. It was literally one command to set up everything like it was, with bootloader, all services, ssh keys, user accounts, .... Then reboot, logging in via ssh, set an account password and that's it.

I also started using it on my workstation, but as you mentioned it is a long and tiresome process to convert my current set up to Nix and have everything work.

And another thing that really annoys me is the added time when you want to iterate quickly on a configuration file (e.g. when setting up a new program), as there is always the build process in between and looking up how to use that specific nixos module, and most of the time I need to look at the source code of the nix module to perfectly understand how to use a specific config option. So in the end I need to read the official documentation for that program, then the nix documentian and/or the nixos module definition. I have a powerful machine, but it still takes 20-40 seconds for a one line change, which would normally be 2-3s (editing the config + restarting a service)

> I have a powerful machine, but it still takes 20-40 seconds for a one line change, which would normally be 2-3s (editing the config + restarting a service)

A bit late reply, but by default Nix runs on single core. Have you made it use multi-core and max-jobs?

Thanks for the heads-up, will look into this, I think I tried something with this but didn't see much better performance, but has been a while since I tried that.

Okay I looked it up, and the default is max-jobs=auto (uses # of logical cores) and cores=0 (each job can use all threads), so if I have a machine with 16 cores and 32 threads, it may run 32 jobs at the same time, where each job is allowed to use 32 threads.

See:

https://search.nixos.org/options?channel=unstable&show=nix.s...

and

https://nixos.org/manual/nix/stable/advanced-topics/cores-vs...

My experience is different. I have a Twitter account that I don’t really use, and am a very active GitHub user. One time someone wrote a Twitter mentioning my Twitter handle with a link to my GitHub repo and commented something to the effect of “you should try out nix instead of what you’re doing in this repo.” Since my GitHub and Twitter handles are different, it means that they specifically google my Twitter account and call me out there. Creepy.

I think NixOS users are generally happy with using it but don't want the support burden of explaining it to others on random internet forums when it's much more different from other distros than Arch is

The thing I like most about Thunderbird's new UI is that I can switch back to the old one. The thing I like least about it, is that the switch is not in the View menu, but in it's own, very small button, out of the way, next to the Quick Filter buttons. What a choice!

> What I really think Thunderbird could use is account sync.

NOPE.

For account sync to be reliable and largely pain-free, credentials would also need to be synced. Which is a massive security no-no, as you don’t control the cloud they’re synced across. It’s why I still rock a pre-cloud version of Mailwasher.

Besides, aside from initial set-up, how frequently do your account settings need to be updated? I have run the same Thunderbird setup without any major changes for about 15 years now. A good ⅘ of my accounts in TB are unchanged since initial set-up in 2004.

If any Thunderbird dev is here please bring the "Customise From Address" to the mobile. It's a god-mode feature (probably nowhere else) for battling spam and truly utilising catch-all for personal domain based emails.

PS. Also, any ETA for iOS? Alpha? Beta?

> please bring the "Customise From Address" to the mobile.

It's a very old issue that appears to be terribly misunderstood at this point, unfortunately the issue on github is currently locked https://github.com/thunderbird/thunderbird-android/issues/93...

> locked as off-topic and limited conversation to collaborators

Companies that do this to their github projects want fans of their product to go away and stop bothering them. I wasn't aware that Mozilla was one of them, but it doesn't really surprise me.

Mozilla hasn't been involved in Thunderbird development for years. They just provide legal and financial backing to Thunderbird via a subsidiary named MZLA. Development is all from the community at this point.

Opinions of Mozilla aside; the reason maintainers do this is because people google bugs or "why can't I do this in x" and they land on GitHub issues. They then pepper those issues with either low-tier comments like "zomg +1 op" or totally unrelated comments. GitHub doesn't have a way to differentiate a random Internet user who arrived on a big through an internet search and someone who is a long time user but not a contributor. This, lock to maintainers happens once you know the cause, effect, and what it takes to fix it.

> They then pepper those issues with either low-tier comments like "zomg +1 op" or totally unrelated comments

I agree it's annoying, but you can also just unsubscribe from the issue if you're no longer interested in community feedback.

> This, lock to maintainers happens once you know the cause, effect, and what it takes to fix it.

This is not really my experience. Usually it goes more like this:

1. Person makes issue

2. Lots of people respond to the issue because they experience it as well

3. Developers either don't care, or fail to grasp what the users actually want

4. More people pile on over the course of months/years while no resolution exists

5. Developers get annoyed, and lock the issue

6. Either they already did, or some point later the developers introduce a "close old issues with no response" bot

7. Issue is locked, cannot get responses

8. Issue gets closed

9. Open issue count goes down

10. Yay!

Pretty sure iOS Mail supports this

I cannot set a random outgoing From: address when sending an email in iOS. In many ways I understand applications not wanting this feature. Non-technical people will see it and wonder: What if I put in someone else's email address here?

Sure you can, it's just a pain (you have to pre-add it in Settings).

Pretty sure it doesn't :) Unless you want a million (okay, couple of dozens or maybe early three digits) email addresses pre-added that you can "select" from.

PS. You might want to check that feature out that I explicitly named on Thunderbird desktop.

PPS. General observation: any software (or service) by Apple isn't even close to, and by that I mean within thousands of miles of, competent third party alternatives (FOSS or otherwise) :P

I assume Thunderbird just lets you type whatever you want there, which I concede is nicer. But in a pinch Mail also does this.

Also, while open source alternatives are generally more flexible than what Apple offers, sometimes this is not the case. Believe me, I use both based on whichever one meets my needs :)

Okay, I will try to keep it short.

> also does this

Nope!

> Believe me

No, but I won't - especially considering how you are pretty much stuck on "yes, Apple mail does it too.. well.. yeah.. in way… yes it does."

Love that Thunderbird is moving forward. The UX work is great, and looks like they are finally getting to some more deeply buried issues too, like https://bugzilla.mozilla.org/show_bug.cgi?id=479969 - my main reason why I can't switch fully back yet

Have you checked betterbird? It's thunderbird with qol bug fixes and features.

The page starts off with light text on a dark background, then as you scroll down it switches to dark text on a light background and then as you keep scrolling it switches back again!

It's impressive to find a page which will irritate both the dark-mode and light-mode users in equal measure. Well done, Mozilla.

Since no one's posted it yet, here's what it used to look like

2017: https://archive.ph/YNAQP

2023: https://web.archive.org/web/20230420163425/https://www.thund...

2024: https://www.thunderbird.net/

What a funny experience!

I started with 2024, continuing with 2023, ending with 2017. Each time it took significantly less time to open and render the page. I didn't check why that happened though, might just be because archive.ph has less load the archive.org etc, but it was extremely noticeable

Yes, archive.org is really slow.

Also: live site and archive.org both load the JS, archive.ph does not.

here's all of them at archive.ph:

2024: https://archive.ph/t2az0

2023: https://archive.ph/Zy43I

2017: https://archive.ph/YNAQP

"Make Email Easier" grabs me. "Free Your Inbox" doesn't. What's all this "Freedoms" razzmatazz? I hear that term a lot but usually from a certain corner of the political spectrum that I treat with extreme skepticism.

The new website looks like every startup website I've seen since the start of the 2020s.

And the header links back to the blog, not the main website.

#1 annoying thing for me after landing on a blog... Never figured out why a company would make it so hard to see the actual product for new users. I can't imagine how many potential customers are lost that way.

And the older versions looked like every website of their era too. Back in 2011 with the "floaty" design (as I can't seem to figure out the real design term) and in 2015 with the material 1.0 design. That's normal!

That's how websites usually work especially when they are mostly portals for the actual product, not the product itself ;).

I really like the new design, much better than the sort of material 1.0 flatness it used to have.

Longtime Thunderbird user. I prefer the old interface and so I've selected that. I wish I could compose a message or a reply in a tab. I often have to click away from the compose window and finding it again is less convenient than it could be (I'm on a Mac.) Lots of people want it, it's been an open bug/feature request since 2008 [0] and yet this is ignored in favor of other UX cruft and website PR.

I don't get the importance of announcing a website redesign. Doesn't look like it would convince anyone to try it.

[0] https://bugzilla.mozilla.org/show_bug.cgi?id=449299

I really wish I could donate to the project without giving them my full details. Why if I send you a couple bucks with paypal do I need to provide my address? PayPal does all the payment validation...

Has thunderbird started doing what Firefox does and categorize your emails to send those stats back to Mozilla corp? If so, or it's in their plans, I guess that would be a reason they may want all my data.

I was happy that thunderbird was abandonware, the protocols haven't changed in the decade since I started using it, and it kept Mozilla from bloating it like FF and seeking profit motives.

I may just move back to roundcube.

In this day and age, why does Thunderbird still ask for my password to store locally instead of using a standard OAuth flow? Every time I consider using Thunderbird, I just can't bring myself to enter a password. It feels like such an antiquated violation and gaping security hole.

Thunderbird supports using TLS client certificates or Kerberos as alternatives to a password for IMAP access.

When you do choose to store a password locally, it's stored encrypted using a second password of your choice.

Since the end result of an OAuth login is a "token" (password...) stored on your machine, I think the difference is pretty marginal either way. But I do hear they're working on OAuth-for-IMAP support. If it were more standardized they probably would have implemented it sooner.

I'm not sure what you mean, I just added a new Gmail account to test and it went through the normal OAuth flow. I didn't have to enter any password into Thunderbird itself, just the Gmail OAuth popup.

An other problem I have is when setting a master password, it doesn't use Linux PAM to re-use login session password

On the other hand, I want the Thunderbird master password to be distinct from my login password. It probably should be a configurable option.

Some email providers support oauth. For some of those (Microsoft and Google), Thunderbird does support the oauth flow. However, how you connect oauth to email is not really standardised, so my understanding is even though "it's all oauth", these implementations are a lot more vendor specific than you'd expect.

Also the need for the email client to have a relationship with the oauth provider is probably a discouragement for some of the smaller email providers to move to oauth.

Some providers actually need a password. Not everybody supports OAuth -- Thunderbird does, and can use it when it's available.

As far as I know there are only two email providers that support OAuth: Gmail and Microsoft. Each uses their own standard to do it, but both are fully supported in Thunderbird with no passwords stored locally.

It's too bad that Thunderbird's UI keeps getting gradually worse, with functionality lost, in newer releases. I don't feel comforted by a website revamp.

(And that's not getting into TB's internal dynamics, which are super problematic.)

What functionality are you missing? All the old menu's are available if you really want them as far as I am aware, just hidden be default.

They broke Manually Sort Folders [1] with TB 115.

That's a big downside for me but what can I use instead of Thunderbird? I worked around that issue by sticking some folders to the top of the list using Favorites. It's 1% of MSF functionality but at least I finally used Favorites.

[1] https://github.com/protz/Manually-Sort-Folders/issues/199

One example is the calendar that keeps getting worse release after release, so much that I had to switch to macOS calendar.

Global Inbox also got broken several years ago and never fixed, since it seems all their energy goes into making things prettier at the expense of making things work.

> making things prettier

And even that is a matter of taste. They tend to make things resemble some combination of an Android app aesthetic and a Google Chrome aesthetic. For example, replacing dialogs with item hierarchy trees with a scrollable flat web-form-like preferences page.

At least they backed out of the switch to monochrome folder icons. 8-\

1. There's no longer a proper toolbar for the main window. (You can sorta-kinda get something that's like it - but not quite, and it is extension-hostile, i.e. won't let you overlay it). This is doubly problematic with the menu bar being hidden.

2. No multi-folder selection. This is necessary for actions involving multiple folders - from a simple move to extension-enabled actions like "Search these folders for duplicate messages")

It's disappointing for that there is no FreeBSD release nor any easy way to find the source for that matter.

Even within their github, there's no source of the client unless I'm missing something?

https://github.com/orgs/thunderbird/repositories

Edit: Wow, to the down-voters you seriously hate FreeBSD that much that I cannot even state my opinion without yourselves needing to down-vote? Why?

It is disappointing that we can't have another OS in the same realm as others and yet we have to keep the whole fanboi cliche going. It's ridiculous, we live in 2024 and yet still have this petty rivalry.

If Thunderbird is able create packages for Snap, Debian, CentOS anything else Linux -- why can they not create packages for FreeBSD?

The source is in Mercurial (like most code from Mozilla), not GitHub. There's instructions on developer.thunderbird.net.

Great thanks.

It seems that Thunderbird itself uses mercurial, not git:

https://developer.thunderbird.net/thunderbird-development/ge...

But yeah, it is really weird that there isn’t really a way to find the source code (or this developer.thunderbird.net page) from the main site as far as I can tell

It's on the Contribute->Participate page.

Oh, I was looking at the blog site but assumed it’s the main page because I got there by clicking the logo and it had buttons like “download thunderbird”… my bad

Not your bad at all. It is definitely their bad for having the logo link to the blog instead of the main page.

It’s in FreeBSD ports, what’s wrong with that?

https://www.freshports.org/mail/thunderbird/

Nothing wrong, but why isn't the package listed upon Thunderbird's website, nor created officially when they do so for Linux?

From someone who complain abouy FreeBSD haters you seem very little informed on what is FreeBSD, how the project is organized and how people typically install software on it.

I am well aware that I could do "pkg install thunderbird" or use ports.

My own preference is to explore the repo and compile it from the repo itself. Do you have a problem with that?

>My own preference is to explore the repo and compile it from the repo itself. Do you have a problem with that?

You are shifting the goal post. You don't need a prepackaged FreeBSD release for that, you can pull the sources from mozilla's mercurial repo.

Not at all, reread my post.

> It's disappointing for that there is no FreeBSD release nor any easy way to find the source for that matter.

The point I'm making their were no packages for FreeBSD nor any easy listing of accessing the source.

The GitHub repo I found, did not contain nor was I aware that Mozilla uses Mercurial. Precompiled packages exist for Linux, but not FreeBSD outside of the FreeBSD port system. Not even a mention.

I agree the information about the location of the source code is buried in the developer doc that you access when hitting contribute. I think an open source product could do better and make the source code more easily accessible.

The source code is here: https://hg.mozilla.org/comm-central, although you also need a checkout of https://hg.mozilla.org/mozilla-central for the Gecko code.

Also, Thunderbird does support FreeBSD, in that FreeBSD-specific bugs will be triaged and fixed, although it's obviously not a primary platform of support.

Why should they bother making a specific release for FreeBSD if a port maintainer does the job well?

Why should it be a job of a FreeBSD port maintainer and not Thunderbird maintaining the port?

Because that is the way the linux distros and BSD traditionnally package software and this is the most integrated way for the user who is just a small command away to get the software. Additionally it removes the burden of maintaining tens of operating system packages to the developers who can still decide to themselves be the maintainers if that is their wish or nobody else want to do it. A win-win situation.

Do you prefer the users to look for software, figure out from themselves they are hitting the right website and not a fake clone with same software but with integrated malware, then to download them and install them with a terrible install process involving hitting next next next install while making sure to not let a checkbox checked that will had some adware like in old win9x days?

BSD ports and curated linux software package repositories are the origins/ancestors of the "app" stores for good reasons. The later only add some voting systems and putting the burden of package maintenance on the software developers, remove curation and don't really prevent rogue mainteners to release malicious clones. I prefer the linux/BSD way.

Yeah, I'm just saying that maybe Thunderbird should create their own package to ensure 100% is as they want it.

Ports only gives myself a selection chosen by the packager, for example Chromium package does not have PulseAudio enabled. I had to recompile it myself to enable Pulse.

> Yeah, I'm just saying that maybe Thunderbird should create their own package to ensure 100% is as they want it.

That was not really the tone of your original comment.

They have perfectly the right to not care enough to not care enough to package it themselves. I am fine if they are just willing to correct bugs and/or upstream any patch a freebsd port maintainer would send them.

Besides you can change ports compilation options with `make config` as well as forking the ports into your own local repo on which you would maintain your own patches if needed. That can be done without any work from the thunderbird or FreeBSD teams.

Either or. No listing of FreeBSD.

Reply on news.ycombinator.com