Hacker News Reader

Comment by tankenmate | Hacker News Reader

I have no direct understanding but I have a hypothesis;

- different generated html / js depending on user agent (UA)

- bug fixes are rolled out on a UA by UA basis

- bug fixes have been rolled out for Windows and Mac (supported)

  - but not for the "generic" HTML / JS output

If this is true it will mean that anything other than the supported platforms will fall further behind as time goes by.

View on HN

That sounds highly likely. Intentional throttling doesn't help Microsoft in this case. I don't think they view Linux desktops (aside from Chromebooks) as a serious threat. Why wouldn't they want more OneDrive users?

This reminds me of the article about the OneDrive team being allowed to introduce ads into Explorer because they don't care at all about the experience of using Windows and they just want more OneDrive users.

(Supporting documentation of the position "why wouldn't they want more OneDrive users"...)

https://news.ycombinator.com/item?id=13903701

A company has an IT director who upon review has determined that Linux workstations would be acceptable for some of his users now due to the fact he uses onedrive.

He tests this solution, finds it slow and decides to stick with Microsoft.

This is probably a happy accident but one they won't move quickly to fix.

Throttling does help if they can get away with it. Now where did I put my windows phone...

That sounds exactly like Google, and makes the most sense.

If this happens to also drive people to Windows, that’s just a side bonus.

Regarding Google: Visit google.com with Firefox Mobile UA, and Chrome Mobile UA. The version presented to Firefox is the Google 2010 version, basically unchanged, with no search options or previews. The Chrome Mobile version actually works.

This is part of why Firefox Mobile now supports all -webkit- properties, webkit’s extensions of JS partially, and emulates Chrome’s UA for a great deal of websites.

> depending on user agent (UA)

This would be very amateurish. The right way to do it is to check browser capabilities. User agents can be easily manipulated and are considered unreliable.

And not every browser bug is detectable via feature detection. I've seen plenty of render specific bugs that were only able to be targeted via user agent parsing. Not it isn't perfect... but it's often the better of the options available.

I'd really love to see what their excuse is.

They're probably doing broad bucketing for supported browsers first, instead of assuming it's supported then limiting/targeting specific versions for various fixes, or non-support.

Hey why doesn't someone just check the sourcecode? ...

Reply on news.ycombinator.com