new interfaces in the works
Here are some screenshots of a new blog page, and a new home page. Feel free to bitch slap my layouts to heathen hell :p oh and if you look closely, you'll see a cool new feature :)
posted by Ali :p @ 3:00 PM
17 comments
17 Comments:
wait... does that include google maps somehow?
yup thats the gooogle map API being used.
Are you going to provide the location data in some machine-parsable format? If so it might be interesting to try and make a Google Earth overlay of the same data.
we don't have plans yet to feed out location data for users, since it is a privacy issue, but we do plan on doing machine readable, universal formats. We'll probably be using longitude/lattitude, so it should be easy to overlay on google earth or other mapping systems.
That's looks hot.
One suggestion. There needs to be not only a way to view all blogs you've STARTED, but also a way to list all blogs you've POSTED in. "View my Blogs" vs "View my Posts".
Make sense?
Also maybe you could add a way to filter out all blogs that were started anonymously. They seem to be mostly on topics of "ill repute" :P
That would be cool too. lol. And you're right. The anon blogs are all "should I abort my baby" or whatever. Bleh!
you said u wont report users locations right? cuz i have a lot of friends on xuqa but i dont want even 1% of them knowing where i live...
Oh, and regards to the "popular" people at Xuqa... lets not only show how many profile views, but also how many scraps, and perhaps kisses/tickles. Will give more variety on the homepage :)
@Anon: Putting your location on the map is more than likely optional.
noah - good suggestion. we're putting in some ways to separate "ill repute" blogs from the ok ones this weekend.
we initially had the view my posts and view my blogs as a single feature, but it started putting too much load on the database. people used it frequently, and b/c the results are updated & unique for each user, and there are like ~500,000 users, its hard to cache results to prevent load on the database. each time anyone clicked on that, we had to run a query searching all the blogs for that users posts and comments. that query actually brought down the site a couple of times.
any ideas on how we could get around it ?
You could qualify the searches to be something like "My posts in the last 7 days". If you factor out the blogs into multiple tables, one per month or something, you can keep the search and index issues a little bit under control. In general it seems that the use pattern of the blogs makes the older data not needed as much. You could keep the index in the per-user data, with the re-indexing done as a low-priority query during off-hours, just looking over the past 24 hours. This would limit the needed query on each click to only the recent (i.e. todays) posts. Without knowing more about how you have the DB layed out its hard to tell how to optimize it, but if you ever need someone to bounce an idea off of, my various screen names and emails are on my xuqa page (http://rpi-us.xuqa.com/Noah_Kantrowitz).
noah - i think that's a nice way of doing it, you're right about really old material not being useful in blogs. we'll work on it in the next upload.
as for this upload, sorry we're running late guys. have a lot of backend changes that require serious QA-ing. we'd rather go a few days late on the upload then put up something that may be buggy.
OOC, have you guys thought of putting up a public bugtracker? Might help with some of the post-release bug squashing.
@Ali: You just need to do what phpbb does when searching. Instead of searching "every single post" for a specific word or author... the phpbb team instead creates a "search word table" which is much SMALLER than the actual posts table... and searches the word table instead.
So when someone starts a new blog, or posts to a new blog, you have 2 fields in a table: user_id, and blog_id
16
1038,1089,2199,3145
The second set of numbers are the blog_ID's that I've posted to. Now when I click on "View My Posts", it simply grabs the blog_id's saved in the new database, then grabs the blog title to display for the user. Clicking on the blog link would then load the entire blog.
Does that make sense at all?
Noah (The other one): The issue with that becomes keeping the cache consistent. I doubt keeping all blog data forever is a viable option, so then whenever an entry times out, you have to search through those tables looking to see if it is mentioned. You could also do the check on access, but the overhead on that would be prohibitive.
Xuqa needs to hurry up and get back to normal. I have messages I need to check.
Post a Comment
<< Home