This Summer, we mainly continued some work initiated during previous Cycles. In that post, you’ll see the results of our efforts. We hope you’ll enjoy reading this article!
Import map… everywhere!
During the previous Cycle, we had started our transition from Webpacker to Import map. Introducing it on our public pages only allowed us to understand it better (usage, limitations, drawbacks, etc.). With this knowledge, we felt confident enough to push that change to our dashboard. It’s now in production, and the removal of Webpacker is bringing a lot of benefits:
- there is no need for Node, Yarn, or any other JS dependencies during the build process,
- it allows lazy-loading of status assets that aren’t required on every page,
- it takes advantage of HTTP/2 multiplexing for asset downloading.
Now, time to focus on the page load performance 🐎
Handlebars phase-out
During the previous Cycle, we also initiated the removal of any kind of client-side templating, thanks to Turbo. As we were successful with this change on the public pages, we also pushed the change to our dashboard, and now hunter.io runs entirely without Handlebars. Again a nice benefit on many levels!
A French dashboard
French-speaking Hunter users probably have already noticed it: Hunter is now fully available in French: Bonjour !
This is a significant milestone that will more languages will follow now that an entire pipeline to support internationalization in our application.
Natural Language Processing in action
Finally, we worked hard on the first usage of data extracted from Natural Language Processing. That is, in short, the ability to extract data from raw text and pages that don’t even contain any email address. For Hunter, that is a game changer! While being extremely powerful, it also comes with some drawbacks. This is why we’re going step-by-step, with a cautious and meticulous approach.
Though, we are happy to say that this kind of data point is now used in our application to improve the precision of scores (Email Finder and Email Verifier) and the quantity of data (Domain Search). Finally, source URLs from NLP will appear soon in the results. And that’s just the beginning!
These are our most noticeable deliveries for this Summer. As always, feel free to contact us for any questions related to our processes (we use Shape Up from Basecamp), our tools (we organize ourselves with Linear) or our company.
And don’t forget to check our Careers page if you would like to join us: we currently have multiple open positions 🔥