| This 39 message thread spans 2 pages: < < 39 ( 1  ) || |
|Constructive criticism WebmasterWorld|
| 8:33 am on Sep 14, 2007 (gmt 0)|
Without getting into a slanging match what would you suggest would be an upgrade for this board? Times have moved on but with all due respect to Mr Tabke WebmasterWorld has not.
For me it lacks:
More modern layout ( get a designer in)
allow links (no follow), but still with moderation. Hyperthetical is better with real life examples
| 12:16 pm on Sep 19, 2007 (gmt 0)|
> Move all of the data into a MySQL database
Perl sql (DBI) routines suck and are slow. To make it work would require moving all the code to PHP. What are you doing for the next two years?
Yes, threads degrade quickly after much quoting is started. You can easily kill a thread on vb by going nuts in a thread with quotes.
There will be some changes available. All optional. What we are trying to do is support the PHPbb template system. That can provide instant access to several thousand templates for users to use here.
> basic is what it is all about
You logged out and looked at a thread recently?
| 2:58 am on Sep 21, 2007 (gmt 0)|
Surprised smilies are still coming up as they've been well supported here for a while :)
I can only say I'm glad to see the poll results of the current poll.
As for quotes, typing [quote] and copying and pasting yourself is a nice way to limit their use. The only thing which does seem missing is attribution for the quote. Were a 'reply with quote' button to be added I'd hope that it would be putting "in reply to <link>" rather than actually repeating the previous content... seen that elsewhere and it works well.
Since it's ask for fixing things season, my list relates to things which shouldn't happen within [code] section:-
|Smileys shouldn't get inserted: |
[[/1]code]print "Hello ".($membername):[/1]($membername)?("guest");[/code[/1]]
print "Hello ".($membername):($membername)?("guest");
|? marks shouldn't have leading whitespace scrubbed: |
[[/1]code]preg_match("/  [/1]?/is","",$string);[/code[/1]]
I appreciate that smileys can be disabled for the whole post; but the setting gets lost on edit and is easy to forget
|Pipes ¦ shouldn't get broken... (although used as field delimiters, would be nice to have them changed back on output due to being key parts of many programming languages - perhaps changing them to | and allowing that through unescaped would be a solution) |
|Indented code support would be brilliant. I know it's bad for small browsers and the like. Not sure how you get around that. A limit of 3/4 spaces used for indenting might allow clarifying structure/loops without being excessive. |
Final attempt to remove broken posts...
| 1:06 am on Sep 23, 2007 (gmt 0)|
|Perl sql (DBI) routines suck and are slow. To make it work would require moving all the code to PHP. What are you doing for the next two years? |
What about mirroring the data in MySQL and only using PHP for search functionalities? Wouldn't even require being done all at once... since it's an upgrade from having nothing at all, you could design the db, code the engine, and port the data in chunks, stepping backward in time. That would make recent posts searchable before older ones, which would allow for checking on industry news that might be getting discussed, without having to wait for any of the se's to spider, index, and publish the pages (which is the biggest pita concerning the lack of search imo, aside from sorting by date).
How is the data being stored now?
[edited by: Brett_Tabke at 2:17 pm (utc) on Sep. 23, 2007]
[edit reason] see sticky [/edit]
| 2:13 pm on Sep 23, 2007 (gmt 0)|
show me and off-the-shelf php/*sql search engine that will work on 10m pages.
| 3:10 pm on Sep 23, 2007 (gmt 0)|
I have another suggestion. When a notification to a subscribed thread arrives by email and the link is clicked we are taken to the first page of the thread. This should by default be the last page since this is where any new post is located.
| 3:15 pm on Sep 23, 2007 (gmt 0)|
|show me and off-the-shelf php/*sql search engine that will work on 10m pages. |
Doesn't have to. Again, since there's nothing there now, anything is really an upgrade... force them to restrict the searches by date with a maximum window of 1 year. Would that work?
Or, if you want both exposure and free labor, since WebmasterWorld is an industry standard, start the WebmasterWorld-Indexation project, the open source effort to develop a PHP/MySQL engine capable of indexing and searching massive legacy sites. :)
| 1:29 am on Sep 24, 2007 (gmt 0)|
>>force them to restrict the searches by date with a maximum window of 1 year. Would that work?
Absolutely not, beyond a shadow of a doubt. I hope you don't mean not being able to search any further back than the current one-year period of time, because there are tons of answers to queries that date as far back as 2002 (or older). How can you specify a date when searching about a mod_rewrite issue? How can you specify a date for a question about CSS for nested unordered lists within a <div>?
And if it's only a one-year window of time for any time period, is that January to January or July to July? And if it takes multiple searches, year by year, here's a biggie: what's the effect on bandwidth usage?
I can find whatever I'm looking for, sometimes as far back as 2000-01, but then again I just do my query in the Google toolbar, which gives me the results in a second or two.
Of course, not everyone wants to run the Google toolbar; but for those who have a problem with searching WebmasterWorld and find it inconvenient to have to go to the search engines to look, would a WebmasterWorld toolbar with a built-in Google search (with some handy hot-buttons) be a more acceptable alternative?
| 3:44 am on Sep 24, 2007 (gmt 0)|
|> Search |
Show me one that will work and we will use it. We have tried every major and most minor packages available on the net. The only one that is even remotely promising (on 5m+ pages) is ASPseek and tests have been going on for awhile. There simply are no software site search engines available that can handle 5m-10m (plan for future) pages.
10 million pages is a significant amount of data to index. The current forum looks to be more around the 1 million mark (looking at only pages that would be interesting to index - eg. the number of threads as opposed to the number of posts). While it is understandable that you may want "future proofing"; be aware that once a certain range is exceeded, it will cease to be cost effective, and unless you are prepared for the cost of investing in additional hardware and/or implementation of a distributed search system, it becomes a less feasible requirement. That is the reason why you cannot find any off-the-shelf product that offers this, rather than anything else.
Technical limitations exist that makes searching 500,000 pages a very different problem to searching millions of pages. For example, available physical RAM, disk space, 32-bit addressing and 2GB virtual address space limit (on 32-bit operating systems) all become issues around the million page mark.
ASPSeek only claims to index "several million pages". There is a reason this is so vague and a "5m+" number is not stated as fact - because this number can vary significantly depending on the content being indexed. You may find that it only allows you to index 1 million pages when it is indexing your particular site, or a lot less. It's hard to tell. The number of words that appear on "a page" varies from site to site, and the nature of such words (whether they are using words of a limited vocabulary, or if they are vastly unique such as in a serial number database), will also greatly affect the resources required.
I work on a software search engine package (which I will omit details here to avoid advertising). In fact, I sent you a private message, but didn't get a response. I was invited here by one of our users who is a member of these forums ... although I've lurked here before, but only thought it relevant to post now. I'd be happy to discuss this in further details if you like.
[edited by: RayL at 3:48 am (utc) on Sep. 24, 2007]
| 7:34 am on Sep 24, 2007 (gmt 0)|
|Absolutely not, beyond a shadow of a doubt. I hope you don't mean not being able to search any further back than the current one-year period of time, because there are tons of answers to queries that date as far back as 2002 (or older). How can you specify a date when searching about a mod_rewrite issue? How can you specify a date for a question about CSS for nested unordered lists within a <div>? |
I was referring to 1 year at a time, not to restricting within a specific year. Besides, just because a system existed for searching the forum, doesn't mean that you personally (or anyone else) would be required to use that system for your own searches... of course you would still have the major SE's to fall back on.
Also, I need to mention this... this thread is on 3450467.htm... is that indicative of it being the 3,450,467 thread started? If so you should realize how much content you currently have that is simply not searchable. Google is declaring a total of 477,000 indexed pages for this site, and only about 169,000 of those are in the non-supp index. The ones that are in supplemental won't be searchable by exact unique quotes. Yahoo is worse at 221,887 pages, and MSN is barely worth mentioning at 46,024 pages.
Marcia, on Jan 1 2001, in one of the threads I see the 677,362th post was made. Yet according to Google, only 2,450 pages are indexed for the year 2000. Is it really meaningful to be able to search such a small portion of the pages from so long ago? And yes, I know the total pages returned by the site: search is just an estimate.
| This 39 message thread spans 2 pages: < < 39 ( 1  ) |