On 18/05/18 12:09, Ken M wrote: > In all honesty I wasn't thinking of the suggestion as a cautious one because > of > bloat. I think bootstrap minified and compressed is like 20k. I mean how big > is > the entire man page collection?
20kB may not sound like a lot, but consider, say, the `sh` man page: http://man.openbsd.org/sh The man page itself (the file in /usr/share/man) is about 53kB in size. There are bigger man pages; some over 1MB (perltoc for example), I picked that one arbitrarily. Rendered as HTML; Firefox tells me that link and all its resources is about 153kB. Throwing bootstrap in there adds an extra 13% to the total page download size (173/153 =~ 1.13). The only saving grace here is that the browser ought to be caching that shared resource, and in ordinary circumstances it will, but you're at the mercy of what the browser (and intermediate proxies) decide to cache. If they decide to not cache: you'll be hit with that extra 20kB on each page view. It doesn't sound like a lot, but it adds up. Back when I was learning web design (did a few-week intensive course back in 2001), I recall the recommendations to design for 800×600 screen resolutions, and to keep load times below 30 seconds when viewed on a 28.8kbps modem link. I don't know what they teach today, but I don't think today's widely available broadband links and high-resolution displays are a good excuse to throw the kitchen sink at the problem. More code means more can go wrong. I find it refreshing to hit an OpenBSD website page, and have it load nearly instantly, and work on any browser released in the last 25~30 years, and without exposing the user to risks such as CDNs going out of business and getting taken over¹. Removing a problematic line from the CSS sounds like a more robust solution than adding unnecessary complexity to the page. Regards, -- Stuart Longland (aka Redhatter, VK4MSL) I haven't lost my mind... ...it's backed up on a tape somewhere. 1. https://blog.hboeck.de/archives/889-Abandoned-Domain-Takeover-as-a-Web-Security-Risk.html