What practical benefit does that offer over existing (synonym-aware) keyword and phrase search approaches? The corpus of one’s mailbox is too small a dataset to draw conclusion from, surely?
The really difficult parts about keyword search are tokenization, normalization, and synonym selection. Especially keeping them up to date. The uses of search never know about any of this, but as the developer these things need to be top of mind.
Using embeddings basically lets the AI configure those things for you and auto updates when the AI updates.
You could also use the embeddings for far more advanced things like in LLMs, but the basic version that is just “better keyword search” is also valuable.
> Not to mention being far slower to query.
KNN on the embeddings is not obviously slower to query. In production using AWS ElasticSearch, for a very large search index, my team saw no meaningful change to latency when using embeddings instead.
If this is for my personal e-mail, then I'd only use it if it could run locally - which also means inobtrusively in the background, not slowing down my computer - is that feasible?
Thunderbird is a fantastic email client, and I'm glad that it has reached a sustainable point. I have worked with plugins just after the switch to mailextensions, and it's great to see that the community has been considered with the changes they have had to make.
However, I am disappointed that TNEF decoding has been removed from the roadmap. The plugin I have been maintaining for this purpose is very old and in desperate need of a rewrite. Honestly, it just needs to be in the core of the client at this point. It is unclear why Microsoft has not completely removed their proprietary mail format, and why admins keep switching it on for outside users is anyone's guess.
Yes, but for me it's more than a client. It's what the "cloud" future should look like everywhere.
If someone provided replicated imap stores, or even better Thunderbird supported replicating the imap store to a one or more encrypted backups, it would be heaven. Access is fast because it's cached locally, but available everywhere and safe from things like "Google locks you out of your account" because it's the cloud.
Thunderbird could do it now with email, calendar and notes. If we could extend that to a password store (like bitwarden), and document store (like Google docs), and a file store (like dropbox), we would be there in the open source world. It's a dream of mine.
An average of $21 is very impressive considering Thunderbird's position. I believe Wikimedia averages to about $15. Although I suppose its position is probably a large part of the comfort to donate more.
I'd be interested to know what their median donation is, assuming $21 is the mean (which it seems to be). I wonder to what extent that number is inflated by a few big donors (it might not be, I'm just curious).
There is a new response from Jason Evangelho about this:
> -Our top 3 donors last year gave 3000 EUR, 2500 EUR, and several that gave 1000 (in both Euros and US dollars)
> -We had ~76 donations last year that were $500 USD equivalent or above (these were made by individuals or small businesses that use Thunderbird, and totals less than 1% of donations)
> -Donations higher than $100 USD (or equivalent) were only 5% of total donations.
> -43% of donations in 2022 were between $5 – $20 USD.
> In our view, we don’t have mega-donors. We have normal, awesome people
Oh, interesting - this is actually about Thunderbird. I was expecting some AI startup founded by people too young to know there was already a Thunderbird.
I don't particularly hate it (I'm used to it after years and years and the changes have been very progressive, probably) except for the change in the "Get Mail" button. It used to be a big clear button with text, and now it's just a small arrow.
Old: https://assets-prod.sumo.prod.webservices.mozgcp.net/media/uploads/gallery/images/2021-10-11-13-20-13-5076f8.png
New (upper left corner) : https://149677182.v2.pressablecdn.com/wp-content/uploads/2022/06/thundebrird-planned-redesign.jpg
Looked at the staff list for a product manager, or a UX designer ... eventually found "product and business development manager" (and "product design manager" and almost everyone else is an engineer. I'm guessing the product and design people are super overstretched and don't have a lot of sway. So that would explain why the UX is what it is.
It's quite functional but the UX is poor. I had to Google how to add a second account. That should be very intuitive to locate for a newbie. It feels very much like other FOSS: built by developers, for developers. I think they could unlock a lot more customers if they employed a designer and product manager.
Yeah, the UX could do with some improvement but IME once UX people have their claws in they change the design every 6 months to give themselves something to do. Anyone who works in the industry surely shares this pain.
Huh, I've always really liked the Thunderbird search; it's really fast, and the clickable date filter is so much easier than any other such UI I've tried. But maybe I search more by date-ranges than most people :)
Agreed on all points, and yet it's still terrible. It does a 2000s-era "fuzzy search" which is infuriating, and just straight up misses exact matches all the time.
Indeed I end up using the date filter as a fallback. It is definitely fast so I can't live without it.
What aspects of the UX do you think that needs improvement? For me the default layout (sidebar for email accounts and folders, upper frame for email list and lower frame for single email) looks very good. The pain is in finding the setting that I need to adjust and specifying criteria for searching emails. However if the improvement is only about making it look better, then I wouldn't buy it because "modern" designs consume more screen space and GPU. Better to have Hacker News than Discourse.
How odd. I think Thunderbird's search is it's most amazing feature.
Maybe that's because I'm a programmer. It can do a multi keyword stem search on literally gigabytes of emails, and return the results in 10's of milliseconds. It looks instantaneous. It's genuinely shocking to me how fast it's inverted index is.
And ... the results are accurate. It's so fast I sometimes have doubted it, and checked the result with grep. It's always right - and grep is much slower. Outlook in comparison, which is one of the few other desktop email clients remaining standing, is just terrible - it's search is slow, inaccurate and so feature bloated you never can be quite sure what corpus it's considering. And it corrupts itself regularly. But I guess comparing it to Outlook is setting a very low bar.
im curious which features you find powerful? as email clients go, it seems to be almost as basic as it comes, with many before it having far far more features.
This is good news to me. I gave Thunderbird another go just last week, as I wanted a way to reliably get my email OUT of MS Office (work), Proton, Gmail, etc. so I could have a local copy in a format (or formats) that I choose.
After I downloaded everything I ran `notmuch` on the lot. My end goal is to have some kind of locally hosted knowledgebase with nearly-instantaneous full-text search. Haven't figured out what comes next, but I'm grateful that Thunderbird made it pretty painless to get this far.
That's also what I use Thunderbird for. Taking control of my emails and importing/exporting them between email providers. I wonder if there is a version of such software that can run on a cloud server though.
I commented in a recent thread about GitHub at last being able to search for "WORD" and finding "-DWORD" in results, they took somewhere around 6 years since I had the need, so it was a welcome improvement... so maybe 2023 will be "the year of search on the desktop" and Thunderbird will start improving its UX for searching emails? Who knows.
Meanwhile, it's been my go-to email client for ages, despite this shortcoming which I had to learn to live with. The UI for searching past emails is very bad, so ironically I ended up changing habits to avoid having to search at all (not the ideal outcome of using a tool, but here we are)
Because I usually know which folder contains what I'm searching for, 98% of the time I'm able to find what I'm looking for quickly using the filter bar. This ends up being a much better experience for me than the search tool.
Because the filter bar doesn't put you in a new UI, you keep the benefits of continuing to use list/message panes: (a) UI context not lost, (b) the matches shows as a list, not the hard-to-scan context results that are default in search, (c) clicking on matches shows the email in the message panel, allowing you to quickly find the right message. The filter bar also lets you quickly toggle if your filter term is the recipient, sender, subject line, or message body.
I wish the UI design of the filter bar could be the basic search tool, expanded in some way to allow for search within multiple/all folders, and slightly more complex searches (specify filters for person AND subject line). Perhaps when you click on the profile in the mailbox panel, the filter bar continues to show, letting you search on all folders. If in a parent folder, there could be a checkbox to show/filter all messages in child folders.
The existing search UI could still exist and be used for advanced searching.
> I ended up changing habits to avoid having to search at all (not the ideal outcome of using a tool, but here we are)
I may have gone to an even greater extreme. I installed a webmail front-end for the sole purpose of searching my emails. I use TBird for everything else.
While we are complaining about search-related issues: I constantly end up with messages that I know I have received, that the search might or might not list, but that then can not be opened until I run "repair folder". Happens with multiple accounts with different providers.
Still better than all alternatives I have tried though :D
That only works in a single folder. Yes, I have a huge inbox and love quick filter for it, but the global search hits all my accounts and archives and for all it's faults, is very fast.
I was about to say, if this is _too_ successful, Thunderbird may find themselves with a target on their back, so Mozilla doesn't have to keep fielding questions from users about how they can fund Firefox directly instead of general donations.
I've been disliking Firefox since before it was cool (the last versions I really loved were somewhere in the 1.x series, and I used to set my theme on later versions to something in the 0.x range, though I forget exactly which one)
You cannot donate to Firefox. Firefox is developed by the Mozilla Corporation, which is owned by the Mozilla Foundation. The Mozilla Corporation shares profits with the Foundation, so that the Foundation can advance its goals.
Enterprise features. Want to grant edit access to an online document for specific authenticated viewers using a particular matrix of browser and OS? Click share, edit permissions, send. Want to see if someone is on vacation? Put their name into the "To:" list, it'll show you their current OOO message so you can skip composing the mail. Want to grant read-only access to a particular group of users from a particular part of an AD forest to a specific shared mailbox? Couple clicks and you're done. OIDC auth with MFA so users can use a 3rd party service to index and access mail and files? Click, click, click.
Well it's integration, improved UX, and features that no other product has. Most companies don't have the time and money to implement all the feature requests businesses come up with, but that's literally Microsoft's bread and butter. You can do more with their stack than with anyone else's, guaranteed. (As long as you are using their stack)
I’m always amazed at the staff numbers for these kinds of projects. I’m glad things are going well for them, but that really seems like a high employee count for such a simple project.
Thunderbird has a ridiculous amount of plumbing work to be done just to keep operational even without new features. Keeping up with the Mozilla codebase, on which Thunderbird is completely dependent, but at the same time, Mozilla no longer cares about Thunderbird’s needs, is probably double digit devs effort in itself.
"Mozilla no longer cares about Thunderbird" is not fully true. Gecko developers routinely take into account the impact on TB when doing changes / refactorings.
Meta has a number of apps e.g. Instagram and services e.g. Workplace, hundred million times more users, revenue in the tens of billions and a highly labour intensive content moderation system.
The comparison is silly. Far better would be with mobile app developers who typically have much smaller teams than 24.
And so I would agree that their headcount seems excessive for what is being delivered.
Thunderbird has at least 20 million users. You think meta has trillions of customers? Those outsourced content mod teams are not included in that headcount either
In account terms a for profit company receiving a donated asset will record the donation as a debit to "Fixed Asset" and a credit to "Contribution Revenue.
For non profit:
When recording journal entries, it is important to determine if the donated asset just passes from the organization to another person, such as donating clothes to the Goodwill, which pass to whoever receives the clothes from the goodwill. If this amount is not substantial, then the asset is not recorded; if the donation is substantial then the company records the assets. Substantial means it will have a meaningful impact on the financial statement.
If the asset passes as such and is substantial, then debit "Expense" and credit "Unrestricted Contributions." If someone donates a substantial amount but the asset is not passed to another person, then record the donation by debiting "Assets" and crediting "Contributions."
If the company sells the donated asset above its fair market value, then the amount over the fair market value is revenue. For example, a nonprofit auctions off a trip to Hawaii that has a fair value of $10,000. The trip auctions for $15,000. The $5,000 difference is revenue.
Daily Thunderbird user for about 17 years. Has always worked well. The recent UI improvements are nice (on the eyes), but, I’m so used to the UI/UX, it just blends into the background.
After the browser took off, ignoring and missing things became the main objectives of Mozilla.
Flush with Google money, the voices of highly technical people with a mission were overpowered by the voices of professional non-technical people with an agenda.
I'm not even getting into the woke stuff, but Mozilla saw themselves as the "Open [insert new stuff]" stewards. They saw companies constantly building moats and ( understandably ) wanted to create an "open" alternative:
Remember the phone OS stuff? Pocket? Servo? Rust managed to survive before the axe came. Raindrop, VR stuff, even IoT..
Since they lost almost all relevance in the browser "market", management somehow "managed" to re-frame their identity to a company that creates cool projects about the latest trend and then lets them rot since they have no traction and eventually closes them.
Google does that too, but Google makes billions of dollars and can support all that "exploration", Mozilla doesn't and it will rot away until they close operations.
I use Firefox since the Phoenix days, and I like it btw.
> We began 2022 with 15 core staff, and now employ a team of 24 in these roles:
Product and Business Development Manager
Director of Operations
Product Design Manager
Engineering Manager
Staff Engineers (3)
Sr SW Engineers (2)
Sr Security Engineer
SW Engineer, Add-Ons Ecosystem
Sr UI/UX Developer
UI/UX Developers (2)
Android Project Lead
Android Developer (1)
Build & Release Engineers (2)
Full Stack Developer
Front End Developer
Community Manager
Marketing Manager
Bug Triager
Support Engineer
I've never seen "Bug Triager" listed as an advertised role. Isn't this a subset of what a product owner/manager typically does?
I tried using Evolution a few years ago, but repeatedly hit problems where if I deleted text while composing an email, "Undo" would not restore it. I filed a bug, which was eventually closed with "we've changed some things around that might have affected this issue, please try to reproduce when the next version is released".
Looking at their issue tracker[1], this kind of thing seems to be a common complaint, rarely resolved with an actual direct bug fix. Most charitably, I assume that Evolution is designed for a use-case or workflow I don't have.
That delete bug doesn't happen for me with current evolution.
The biggest problem I have is the composer window often isn't exactly the same as the sent mail (especially for plain text mail) when it comes to whitespace and quoting. Enabling the Outbox feature is a good workaround as it lets you see the mail before sending it, and you can adjust the mail until it is correct.
Evolution is kind of a web client (like Thunderbird) too, since the message preview pane, including the parts outside the message body, is rendered by WebKit, the HTML for it is generated by evolution. Also the plain text messages go through HTML and WebKit too. Would be nice if these weren't the case though, because they necessarily aren't great for security.
I use bower (fed with isync) in addition to Evolution (because near-instant searching of 15 years of e-mails is rather handy). My coworkers always send email that is readable in a terminal, but apparently customers have this thing called "money" which means they get to dictate whether or not I use HTML e-mail...
I love it. I previously used Sylpheed (and Claws, a fork of Sylpheed), but eventually got to the point where I had to reply to HTML e-mail and that moved me back to Evolution. In addition, its Exchange backend is excellent, which is good because I can't convince O365 to send me calendar invites over IMAP.
Apple's Mail app is constantly losing it's connection to my various different work email accounts, but the annoying thing isn't that it's losing it, it's that it's not doing a good job of either telling me, or having me re-auth. I should probably give Thunderbird another try - specially since they're funding is up.
Yay! I haven't used Thunderbird much in a few years, but I'm very glad they're still around. Last I tried it, I was able to pay for a cheap plug-in to authenticate to Outlook365 and retrieve corporate mail through Thunderbird for a Linux laptop pilot for work. Very glad people still have an alternative to the Enterprise Empire.
I like Thunderbird, but on Linux I finally switched to Evolution because Thunderbird was an absolute RAM and CPU hog, and also, Google Calendar integration was flakey as hell. Thunderbird would randomly decide to not notify me of an event or decide to notify me way too late.
I would use Evolution if it didn't look and feel like the version found in early GNOME 2 Ubuntu, when they still used to ship installation CDs for free.
Aesthetics is as important a feature as supporting IMAP or having decent search. Which is why I'm still stuck with Fastmail's webmail, because neither Evolution nor Thunderbird seem to satisfy these criteria.
I like Thunderbird and use it all the time (it is ideal for my workflow with a couple of add-ons installed) but the developers should focus more on the application's performance. It loves to bog down my system.
JMAP doesn’t seem to have wider provider support (Fastmail, which created JMAP, seems to be the only one supporting it). It also has some pieces (IIRC, calendar, contacts) that are still being defined. Thunderbird supporting JMAP isn’t going to do much for Thunderbird.
On the other hand, if Thunderbird supported Microsoft Exchange natively (without extensions) for mails and calendars using EWS (deprecated) or MS Graph, that could increase its use in the enterprise manifold. From what I’ve seen, companies using Microsoft365 have been turning off IMAP (and POP), pushing employees to Outlook as the only option.
Is it all that useful for regular desktop mail client? I can understand it's huge usefulness in browser or on mobile but on the desktop? Yes, it simplifies mail management, but (clunky) IMAP is already and works just fine?
> but (clunky) IMAP is already and works just fine
Well, 'works just fine' is not what I would say about IMAP, more like 'works most of the time'. But in the end, IMAP is just for accessing mailbox and mail. JMAP also covers (besides server side actions and settings) calendar and contacts. If you remember, there are a separate protocols for these now (CalDAV, CardDAV) which are totally unrelated to IMAP. A desktop client with a full JMAP support would be close by the functionality to Outlook connected to Exchange, which is a thing many desire, especially in corporate settings.
I'm aware and use those protocols. What's the benefit of bundling IMAP/SMTP/Cal|CardDAV into single JAMP protocol here? Those separate protocols work great and if someone only needs mail can only use IMAP/SMTP?
Use it out of habit on Windows but been annoyed for years at how slow the UI is. I frequently open an email and start typing a reply and have to wait 10 seconds because it locks up for some unknown reason. I’ve never been able to figure out what’s going on and this has plagued the app for a good decade at least.
Fascinating, per the post, they were struggling in most of 2021 and 2022, and there were some last minute contributions in December 2022 that bumped up their 2022 numbers. Curious what outreach or activity they did in Dec 2022 that led to the influx of funds.
I recently changed employers, and one of the IT policy differences that I'm very happy with is that they do not disable IMAP, so I can use Thunderbird for my work email again.
I basically abandoned all hope of keeping on top of my email at my previous employer when they locked it down to gmail web UI only.
Why does there have to be a Thunderbird at all? The people building Thunderbird are clearly motivated to build FOSS/Libre communication tools for people.
As more and more people move to using Android and ios as their primary computing environment, the existence of accessible FOSS/Libre tools such as Thunderbird becomes more important than ever.
So, if they're interested or dedicated to building open tools for people to use, they can choose where to go, whether you like it or not. :)
e: By the way, have you donated to Thunderbird? If you are able to pay for the software you use, especially when it seeks to be a force for good, then you should. Even $20 a year goes a long way for an organization like this.
An android port is within top 3 requests as per Mozilla Connect's Thunderbird section. Makes sense to me to work on such requests when you're operating off donations.
Now you know and can get into voting for the suggestions you deem valuable!
I don't know if there are any other public surfaces the Thunderbird team uses to figure out what to work on (outside of bug reports/telemetry and Mozilla Connect).
May be it's just me, but I don't use email the same on my phone as I do desktop. I have 5 accounts on my desktop but just the one witch is my mane on my phone. Aps these days cant just stay good at one thing like ff being good on windows. They want to be good on Windows, Linux, BSD, Android so on and so on.
Sometimes I must open some stupid mail on my phone because an app needs it for verification or some other nonsense. I'd like to have an accessible option to do it.
Because it's a real time-waster to need to setup accounts on each device again and again, setup email signatures, setup canned responses, configure as needed etc. Apple has been able to sync accounts and signatures since the original iPhone with their Mail app.
When you follow the link on that page to donate, select monthly recurring donation, and go to enter payment info, they then try to upsell you (with a checkbox checked by default) to "Cover transaction costs".
(When you click the info icon: "When you cover our transaction costs, you're covering both our processing and donation platform costs.")
It's easier for a non-profit charity to be pushy about fund-raising, because it's for a good cause. But that's not Mozilla. Facets of Mozilla are run more like a for-profit tech company, including layoffs of engineers while executives still demand high compensation. Pushy comes across differently.
People want to tip the waiter, and would feel positively about that, at the same time they'd feel negatively about a restaurant owner trying to sneak a made-up fee onto the bill.
I suspect it is an industry trend in donations - I saw this yesterday when donating (which I don’t often do). I think they asked for $17.50 “transaction costs” on a $150 donation.
I wish there was an anonymising service in New Zealand so I could donate without giving my contact details. No, I don’t want to be added to a list - there is a whole seedy underbelly to the industry of charity.
You should probably be directing your outrage at the platforms that are taking those costs, not the non-profit who are effectively forced into paying them.
The biggest shame of the cryptocurrency debacle is that it was never able to make any real headway into the stranglehold that the Visa/Mastercard duopoly has on online payments.
Impressive!