On 7/11/17, Glen via support-seamonkey <[email protected]> wrote: > Paul B. Gallagher wrote: >> Lee wrote: >> >>> On 7/10/17, Paul B. Gallagher <[email protected]> >>> wrote: >>>> Lee wrote: >>>> >>>>> I was really hoping to skip the discussion of exactly what >>>>> "little benefit" means & go straight to how to decide _for >>>>> yourself_ if disabling cache is a Good Idea or no. >>>> >>>> Fine. Try it, and if you like it better, stick with it. If not, >>>> revert. That's how I'd decide. >>> >>> Really?! You wouldn't even try to look at any performance data? >>> >>> Whatever works for you, but I've seen too many instances where >>> people see what they want/expect to see instead of what's really >>> there, so I like looking at the numbers instead of going by feel. >> >> I would happily look at performance data for comparable systems under >> comparable conditions. Got any of that?
Where did performance data for comparable systems come from? I was objecting to this bit: >> If you have a broadband connection there is little benefit to using a >> browser cache. Disable it. > > Interesting and clearly true, "clearly true" is not at all clear me, so I gave a way to check if it was true or not >>for you<<, that was not subjective, didn't require any new software and was easy to do. >> My point was that systems are so >> varied and conditions are so varied that it's impossible to make a fair >> comparison. All we have is empirical results for our own systems. And >> ultimately that's the bottom line -- is this user happier or not? have a look here https://www.theregister.co.uk/2017/07/07/bofh_2017_episode_9/ If you've ever had to do customer support you know it's true :( 'Is the user happier or not' is influenced by a lot of factors, including the expectation that a proposed change for the better influences one to think the change really did make things better. > Paul-1, Lee-0. I'll have to agree with Paul on this one. That you don't even try to get any performance metrics before/after making a change? > For example, > Lee, if you took two identically built computers (same models, specs, > etc) - like Paul was describing - even taking it to the point where both > systems were connected to the same router, in the same house, one could > have loads of software on it, rarely maintained, while the other could > have limited use and constantly maintained, and then did performance > tests on those two computers, well, most likely, your results could vary > so greatly that you'd start to question whether those two computers were > identical models or not. Like Paul suggests, you truly need > apples-to-apples type testing here, not just looking at one computer's > performance test and then draw a conclusion by saying, 'All systems > should make this change to it.' I am amazed you can get that from what I wrote. I'm advocating that people do the test on their system & pick whichever way works best for them. If you go back to my first msg, I was objecting to the 'all systems should make this change' recommendation. I don't know how we got into apples-to-apples testing, user happiness, extrapolating results from one machine to another or whatever. I like having objective measures to see if a change makes things better or not. If you're like me and would rather know than guess - you've got a way to get objective performance numbers. Regards, Lee _______________________________________________ support-seamonkey mailing list [email protected] https://lists.mozilla.org/listinfo/support-seamonkey
