A point of reference

Ben Bucksch ben.bucksch at beonex.com
Wed Mar 7 20:47:11 UTC 2018

Joshua Cranmer 🐧 wrote on 07.03.18 21:27:
> On 3/7/2018 3:19 PM, Ben Bucksch wrote:
>> Most of these are in the form of webapps replacing older native apps.
>> Most users use Gmail instead of Thunderbird and Outlook. That 
>> includes power users that are speed-sensitive. In fact, that's 
>> Thunderbird's main competitor: webmail.
> That's an apples-to-oranges comparison. An email client (to a first 
> approximation) is centered around a core database, and it's the 
> database operations that are performance sensitive.

If indeed the DB turns out to be a bottleneck, there's no problem in 
using a native C library for the database - e.g. sqlite, or something else.

Joshua, if you have specific ideas about how a future Thunderbird 
database should be designed, in API and in implementation, please do 
write your thoughts down (in a new thread, or on the wiki), so that we 
can include them, if we ever get to implement something new for TB:Gecko 
or TB:NG.

> asm.js

OK, point taken. Bad example.

However, there are many speed comparisons between JS and C/C++, and JS 
on a modern JIT doesn't fair badly.

You do make a point that *some* uses of JS are slow. As a simple 
example, jquery uses a lot of strings for everything, including for 
calling functions, which has to be slow.

We also have the option of using TypeScript or something reasonably 
similar to JS, if indeed that performs better.

This is why I advocate of being specific. I don't think a blanket 
statement "JS is slow" is fair or true. Not since V8, and certainly not 
in 2018.

Without more specific measurements and investigation, we're just arguing 
and bickering.


More information about the tb-planning mailing list