sphinx search 2.0.0 nearly completed +
I’m in the final stage to release the new Sphinx product for vBulletin. You will be pleasantly surprised with the new version.
Here it is a little screenshot for one of many new options, in Admin CP. In Sphinx 2.0.0 product (compared to 1.x.x version), you have much better grasp of what is going on behind the search scene.
Or maybe some suggested keywords like Google? 
You will see all details soon, when I finish the release. Read the comments for futher information and updates.
More updates. A major improvement was completed and tested, as we speak:
Instant Content Updates (10min interval)
Right now, the product will update a thread/title or content, once a day. I did not liked that at all. I want to have a thread title or post content updated into searches ASAP, not after 12 hours or whenever is due the major index update. I burned 3 weeks of intense work on this feature, because I find it very important. The scenario is simple:
If an user edits a thread/post within the first 10min (prior initial index), then the data will be stored into search index as usual. The trick is to update the data after the original index was performed. This mostly occurs when a moderator reviews a thread/post and decides to either edit/delete the thread/post title or content.
Sphinx 2.0.0 will allow the previously indexed data to be automatically edited/deleted from indexes within 10min or less, so only the relevant search information will be displayed to users. You can also review easy what threads/posts are scheduled for updates, into Search Log:
The links displayed into Info column will point you to the actual thread of post, so you can easily review the changed thread/post.
The Search Log is extremely useful, because the vast information stored into. It allows an administrator to properly view any errors, warnings and other hidden activities related to Sphinx Search, while he/she is away from the boards. Below, you will see the data types stored into Search log:
I will post more updates, with the work in progress.
Another important feature that I wanted to talk about is the… speed. Who does not like things going fast? Fast cars, fast bikes (I ride a Warrior, pretty damn fast), fast computers and… fast indexes.
In current product, only one processor is used regardless how many you have available. So if you have a dual quad core, only one processor will be used for searching, while the other 7 sit there and do nothing. Not good.
In Sphinx 2.0.0 the search speed was improved because the usage of all processors available. In other words an index can use multiple processors to work faster for you, resulting into even faster results. Not that Sphinx needs it, but still, faster is always better. You can see it in this screenshot:
Status: Serving 8,563,362 records on 6 processors
Below the general information, you can actually see how the large index is split into 6 dynamic indexes, each assigned to a specific processor. You can also see the number of records (threads or posts) and other useful information contained into each index. As I said before, the more information you have on hand, the better you manage your searching data.
When the index grows over a specific number or records (posts), it starts to use an additional processor. When the data grows to surpass the 8 processors capacity, the current index is “re-split”, freeing 2 processors. That explains why only 6 processors are used in the above screenshot, instead of 8. It simply avoids frequent data manipulation for no reason. The dynamic indexes are in “dormant” state, ready to be waked as soon as data is inserted into them. They also get updated only if new data is assigned to each index, resulting in even less I/O usage.
Just to “show off”, I want to give you a better example related to speed improvements.
Right now, to update all indexes into a 10GB database takes in average 30min. Every night around 3am a little cron job starts and makes those disks burn like fire, in order to refresh all data for 30min or more… Do you think this is good? No, it is not good at all…
What if I tell you that in Sphinx 2.0.0 the indexing is 1500% faster? That’s right, there is no typo in the percentage number. In other words, an index build that takes usually 30min it will take in average only 2-3min. Now, that is fast.
Sphinx 2.0.0 for vBulletin is now released. All customers with active licenses get a free technical upgrade.
Why don’t you sell this at a lower rate and attract tons of more buyers ? lets say at $200/copy with limited features (say a lite version) , with no installation. Just put a guide on how to install or just charge extra for installation service.
You can always have a premium copy/version which you can sell at ur current rates.
For the simple reason that my product targets large forums who can afford the price. Also, having this price tag will eliminate any pirating cases, plus will allow me to focus better on customer service. If your board needs my product that means you are certainly making the cash to cover this type of expense. One server cost in average $5,000/year. You do the math and see how much you save.
For your information, I tested several clients, asking them if they are able to install the product. Nobody was capable to complete all tasks properly. I will certainly not lose hundreds of hours on the forums explaining to people how to perform basic server tasks.