This is a repost in plain readable ascii.
I wonder how much web bandwidth and server processing power is consumed by diligent wikipedians browing the list of 'recent changes', checking an article and returning to the list by pressing the back button, thus generating a new request to the database, over and over again. An expensive request as well, in terms of database processing (I presume) and html page size, up to 500 records might be sent.
My suggestion: a very simple change would reduce this load on the server considerably: add target="_blank" to each link on the 'recent changes' page only. Now articles will be shown in a new window. The 'recent changes' list will only be refreshed when the user explicitly asks for it (via refresh button or menu click). This may alleviate slow response times somewhat.
Not posted on the technical mailing list, since functionality will obviously change (for the better in my view, I would like a subsecond redisplay of the changes list).
Erik Zachte
Erik Zachte wrote:
This is a repost in plain readable ascii.
I wonder how much web bandwidth and server processing power is consumed by diligent wikipedians browing the list of 'recent changes', checking an article and returning to the list by pressing the back button, thus generating a new request to the database, over and over again.
They should use a browser that's not so brain-dead that it can't tell that the forward and back buttons are for zipping through pages that are already open and should not be loaded all over again.
My suggestion: a very simple change would reduce this load on the server considerably: add target="_blank" to each link on the 'recent changes' page only. Now articles will be shown in a new window. The 'recent changes' list will only be refreshed when the user explicitly asks for it (via refresh button or menu click). This may alleviate slow response times somewhat.
Please, no. I think I know better than my browser when I want to open a page in a new window.
Not posted on the technical mailing list, since functionality will obviously change (for the better in my view, I would like a subsecond redisplay of the changes list).
Have you considered a browser that provides tabbed browsing?
-- brion vibber (brion @ pobox.com)
Brion VIBBER wrote:
Erik Zachte wrote:
This is a repost in plain readable ascii.
I wonder how much web bandwidth and server processing power is consumed by diligent wikipedians browing the list of 'recent changes', checking an article and returning to the list by pressing the back button, thus generating a new request to the database, over and over again.
They should use a browser that's not so brain-dead that it can't tell that the forward and back buttons are for zipping through pages that are already open and should not be loaded all over again.
not true.
The recent changes page used to remain the same when you went back to it unless you manually told it to reload, but now if you use the back button to go back to a list page it reloads it from the server even if you don't particularly want it to. If I want to save time I have to right-click and say 'open in new window' all the time and it wastes my time and energy. I would be VERY much in favour of that change being made.
And don't tell me that my browser is braindead. It didn't used to do this. The wikipedia was changed, not netscape.
Karen AKA Kajikit wrote:
Brion VIBBER wrote:
They should use a browser that's not so brain-dead that it can't tell that the forward and back buttons are for zipping through pages that are already open and should not be loaded all over again.
not true.
Yes, true. Alas, not everyone lives up to my high standards. ;)
The recent changes page used to remain the same when you went back to it unless you manually told it to reload, but now if you use the back button to go back to a list page it reloads it from the server even if you don't particularly want it to. If I want to save time I have to right-click and say 'open in new window' all the time and it wastes my time and energy. I would be VERY much in favour of that change being made.
And don't tell me that my browser is braindead. It didn't used to do this. The wikipedia was changed, not netscape.
Netscape 4.x being braindead is not mutually exclusive with the wiki's output changing. In fact, it's the conjunction of both that causes the problem.
Currently the wiki sends out a number of cache-suppressing headers which, I believe, were introduced to keep Internet Explorer from showing you the previous version of a page after you save an edit:
Expires: 0 Cache-Control: no-cache Pragma: no-cache
Netscape 4.x and Internet Explorer (at least the 5.5 I have) are so overzealous about these that they force a reload even when using the forward and back buttons. Mozilla (and presumably Netscape 6 and 7) are saner about this, and consider the page still open -- forward and back aren't page load events, they're just showing you things you've already opened again.
I'm not the one that added those, so I don't know exactly what they're meant to battle against, or under what circumstances they are necessary, or how they might be appropriately tweaked to maintain whatever alleged benefit without breaking the angry browsers.
-- brion vibber (brion @ pobox.com)
Brion Vibber wrote in part:
Netscape 4.x and Internet Explorer (at least the 5.5 I have) are so overzealous about these that they force a reload even when using the forward and back buttons. Mozilla (and presumably Netscape 6 and 7) are saner about this, and consider the page still open -- forward and back aren't page load events, they're just showing you things you've already opened again.
Netscape 6 has the same behavior as Netscape 4 and IE 5 (all for Unix). This has annoyed me in other situations as well.
-- Toby
It has nothing to do with how brain dead your browser is, but how its preferences are set. I can set IE to refresh a page every time I come to it, to only refresh once in a session, or to refresh whenever content is change. Zoe Brion VIBBER brion@pobox.com wrote: Erik Zachte wrote:
This is a repost in plain readable ascii.
I wonder how much web bandwidth and server processing power is consumed by diligent wikipedians browing the list of 'recent changes', checking an article and returning to the list by pressing the back button, thus generating a new request to the database, over and over again.
They should use a browser that's not so brain-dead that it can't tell that the forward and back buttons are for zipping through pages that are already open and should not be loaded all over again.
--------------------------------- Do you Yahoo!? HotJobs - Search new jobs daily now
Zoe wrote:
It has nothing to do with how brain dead your browser is, but how its preferences are set. I can set IE to refresh a page every time I come to it, to only refresh once in a session, or to refresh whenever content is change.
Which reminds me: how do they know when the content has changed if they don't download the page every time from the server? What's the practical difference between this and every time that you come?
-- Toby
Good question. I never thought of that. There must be some kind of time stamp they can access? Zoe Toby Bartels toby+wikipedia@math.ucr.edu wrote:Zoe wrote:
It has nothing to do with how brain dead your browser is, but how its preferences are set. I can set IE to refresh a page every time I come to it, to only refresh once in a session, or to refresh whenever content is change.
Which reminds me: how do they know when the content has changed if they don't download the page every time from the server? What's the practical difference between this and every time that you come?
-- Toby _______________________________________________ Wikipedia-l mailing list Wikipedia-l@wikipedia.org http://www.wikipedia.org/mailman/listinfo/wikipedia-l
--------------------------------- Do you Yahoo!? U2 on LAUNCH - Exclusive medley & videos from Greatest Hits CD
wikipedia-l@lists.wikimedia.org