> how is the plan to slowly rewrite it viable?

Last time this topic came up on HN, someone linked to a thread on tb-planning where the post-XUL future was discussed. IIRC the "slow rewrite" was considered unworkable, and the proposed plan was to restart from scratch using "web technologies" while continuing doing basically security updates for the classic XUL TB until the replacement is ready.

As an aside, what "web technologies" exactly meant wasn't discussed. Is it electron? Or something mozilla-based like the servo + browser.html experiment? Or running a separate server process and connect to it using a normal browser (like e.g. mailpile does)?

I would assume along the lines of using writing the core logic in Javascript, using NPM packages to help implement things, and using UI libraries like React, rather than building everything in C++/XPCOM/XUL.

IDK, I like that Thunderbird sorts without conversation. It is the primary feature I look for in a MUA.

The only feature I find annoying is the search feature, which seems to be a universally accepted weakness of Thunderbird.

I'm confused by its two search bars. It doesn't make sense to have two.

One is a global search, the other is a simple filter for the currently opened mail folder. I use both for their respective use cases all the time.

IIRC, the one for the highlighted directory uses regex (I'm really horrible with regex, so I don't use it).

Yeah, the search is really terrible.

I'm thinking when Windows 7 expires in 2020 and I have to migrate to Linux because Microsoft hates having paying customers such as myself, I'll probably move to either Mint or Ubuntu and install a local Sphinx Search server specifically so I can search Thunderbird emails.

Why waiting so long? Move now and give yourself a smoother transition.

The HTML5/JS option is probably the only way out for Thunderbird but it won't seduce current backers. So it's a bit hard for them to say "Thunderbird has been abandoned on an motorway service station".

Well it won't seduce me.

Who wants an email client that uses 600 MB RAM just to open and that eats 100% CPU to animate your email folders?

To be clear, Thunderbird is currently implemented in XUL and JS. Moving to HTML and JS is not going to significantly change its performance characteristics, mostly. There are some details that will need to be done right, like needing to create a good HTML replacement for <xul:tree>, of course.

There comes a time where good sense means you have to kiss these old projects goodbye. It makes sense for Thunderbird to try to move to a platform that users actually want to use. Thunderbird is in the state it's in because very few people still value a "fat" mail client.

Thunderbird supports conversations but they're off by default. But really I prefer the way it works now.

There's also https://addons.mozilla.org/en-US/thunderbird/addon/gmail-con... for those that have become used to seeing the content of all emails in a conversation.

Doesn't work well. I personally like GMail-style conversations, but I couldn't use that plugin.

Yes it's a bit of a crutch. I prefer it to Thunderbolt's default view, though.

> And it should have a modern UI with at least conversation view like Gmail.

I never really found the conversation view of email very useful (especially if there are more then two people participating in the conversation). It's hard to tell who replied to whom and you can't really get context if there are responses to multiple parts of the message they're replying to.

I do have a strong preference for the threaded view that Thunderbird has now.

> And it should have a modern UI with at least conversation view like Gmail.

Gnus nnimap provides a threaded, conversation view. It's almost definitely not what you'd consider a 'modern UI,' but I like it quite a bit.