Florent Bouillon

I'm genuinely kind of confused as to why IPS is posting such an early announcement of 4.4.  While we're all excited for the newest version (insert more gifs here), this really is premature:
1. You can't actually download unless you specifically request beta.  
2. You probably don't want to download since it's beta.  
3.  Did I mention it's still in beta yet?
5. And not even late stage beta.  This is literally beta 1 with potential breaks to global template. 
4.  But whatever.  Party on.  This is like the pre-party before the preview party before the launch party, and by the time we actually get to download 4.4 we will all be so drunk and hungover from the partying that we really won't care what version we're actually using on our live sites. 

I was exactly in this situation a while ago, out-of-date softwares and server on a big community. Just because updating was a lot of efforts, and I always had bettor to do of the little amount of time I had. Then a server crash brutally accelerated the process, and I worked a lot to migrate on some brand new hardware and software, including migrating on Invision. All what you are saying in this article came true: better safety, better SEO, better engagement. Soon the statistics began to grow rapidly after years of stagnation (in particular on mobile), administration was much simpler and time-sparring, and I could enjoy developing much more complicated functions.
So, long live invision 🙂 

I was exactly in this situation a while ago, out-of-date softwares and server on a big community. Just because updating was a lot of efforts, and I always had bettor to do of the little amount of time I had. Then a server crash brutally accelerated the process, and I worked a lot to migrate on some brand new hardware and software, including migrating on Invision. All what you are saying in this article came true: better safety, better SEO, better engagement. Soon the statistics began to grow rapidly after years of stagnation (in particular on mobile), administration was much simpler and time-sparring, and I could enjoy developing much more complicated functions.
So, long live invision 🙂 

I was exactly in this situation a while ago, out-of-date softwares and server on a big community. Just because updating was a lot of efforts, and I always had bettor to do of the little amount of time I had. Then a server crash brutally accelerated the process, and I worked a lot to migrate on some brand new hardware and software, including migrating on Invision. All what you are saying in this article came true: better safety, better SEO, better engagement. Soon the statistics began to grow rapidly after years of stagnation (in particular on mobile), administration was much simpler and time-sparring, and I could enjoy developing much more complicated functions.
So, long live invision 🙂 

This is really good news, I'd been hoping to see a performance-based release for some time as traditionally IPS does rate low in Page SpeedGTMetrix/Yslow tests and we've been waiting for some of these overdue improvements for quite some time. So huge thanks! 
Lazyloading should really help with Gallery in particular, I currently get a poor Pagespeed score of F42 on one of my sites.
Please consider adding to your to do list (if not already added):
Serve scaled images (current GTMetrix reports a score of F0 with a 1.2MiB 80% possible saving is possible on my Gallery index page) Where static content is uploaded to Amazon S3 and Cloudfront (it should be copied, not moved!), often there is missing header metadata such as cache-control, expires etc. Whilst investigating poor Cloudfront cache hit ratios of just 5%, I noticed this week an IPS uploaded .png image placed in my S3 bucket with a content-type of image/jpeg instead of image/png.  Leverage Browser Caching (you can't do anything about Amazon and Google external resources, but IPS could and should add more to improve caching of its static resources) Specify Image Dimensions - I get awful scores for this both on Gallery and Forums where images are missing width and height attributes. This happens particularly in grid view layout and where you upload an image attachment into forum description fields via AdminCP.  Remove query strings from urls - Have raised this several times, and keep getting told it's needed to allow cache busting, but the string token could be built into the actual filename/url itself. If CSS for example is later rebuilt, a new url is created and immediately browsers see the updated resource and recache it. All of the separate CSS and JS files should be grouped into one or two files, not like 10-12. Pages article images should have medium sized thumbnails available instead of just tiny thumbnails which look awful stretch to cover on 4K displays. The alternative is full-sized (in my case massive 1920x1080 images) which kill page loading times are are such overkill for article listing previews.  
 
 
 

Do you think about improving the way the fonts are loaded? It's very unpleasant to watch empty rectangles until ...
Apparently you should also look for a variant of fonts awesome at your choice of administrator, or minified version of the fonts awesome.
This would reduce the size of the downloaded file and accelerate the uploading of fonts awesome.
It would also be a good solution when installing or upgrading, if an administrator chooses which flags to use and this to save a file as a custom_flags.css. This will greatly reduce size the downloaded flag.css file

I don't like current behaviour of highlighting because it works for every posts, does not matter if this post important or not (i.e. "official staff answer" or "staff member's comment on kittens photo"). I'd prefer to have an option below editor. Posted this also as suggestion in feedback and ideas section. Please support(reaction/comment - any activity) this topic if you need it too.