V7 beta release available for testing

[Ray]

scuse me of this reply, But in my case i use the offline index for DVD/CD, I need in the ver 7 the fuzzy search, because sometimes the OCR does not recognize the characters and I can not exactly fix a million pdf to correct them... I think it is not difficult to implement in the engine there are lot of sample scripts on the Internet, to correct the search engine to find "home" and "hpme" is the same word... with 1-2 character of error tnx.

http://en.wikipedia.org/wiki/Approxi...tring_matching

It is easy to match two words with 1-2 character difference. But it is not easy to return good search results with this condition.

If "home" and "hpme" are considered the same word because there is only 1 character difference, then these following words are also considered the same as "home": come, dome, some, tome, hole, hope, hose, ... just to name a few.

I think you have to agree that the search results will contain very irrelevant results when all these words are considered the same.

As previously explained here, Zoom supports fuzzy searches with the same techniques used in most other search engines (e.g. stemming, synonym, metaphones... we even support wildcards which is very rare and make our queries more powerful than most).

What you are asking for is not implemented in any decent search engine because it simply would not deliver good results. Search "hpme" on Google and you will see it is not the same as "home".

Your real problem (that you are trying to avoid) is a bad OCR process, and that's what you should be addressing.

[tommyk]

scuse me of this reply, But in my case i use the offline index for DVD/CD, I need in the ver 7 the fuzzy search, because sometimes the OCR does not recognize the characters and I can not exactly fix a million pdf to correct them... I think it is not difficult to implement in the engine there are lot of sample scripts on the Internet, to correct the search engine to find "home" and "hpme" is the same word... with 1-2 character of error tnx.

http://en.wikipedia.org/wiki/Approxi...tring_matching

[soatley]

I read in the V7 Alpha docs that it installs over V6. Is it possible to install V7 to another folder while still maintaining V6? I'd like to try 7 without removing the V6 functionality from my site.

[David]

When will be ready the v7? If now i buy the v6, after i must re-buy the v7 version?

See my comments earlier in the thread.

Is v7 fully functional with respect to languages? After the install, the lang directory only has English, German and Welsh languages.

Hmm, there should be more than this. We'll check this out.

[pbparker]

Is v7 fully functional with respect to languages? After the install, the lang directory only has English, German and Welsh languages.

[phoenix1977]

When will be ready the v7? If now i buy the v6, after i must re-buy the v7 version?

[Ray]

At this moment, the list of manually entered entries is only 500. It will then take the top x words from your statistics log if specified. You can specify 2000 for the top x words from statistics log, and it will then have a total list of 2500 words to choose from. But you will need to have a statistics log (see this page for more details) and isn't exactly what you want.

We will increase this to 5000 in the next alpha release.

In the long run (before the final V7 release), we will most likely remove the limit all together so that it can be indefinite based on memory, and also pick up words via meta keywords tags during indexing (e.g. use the meta ZOOMWORDS).

[Josiahjg]

We tried to update our list of autocomplete and it said we put too many results in the box (in Configuration > Autocomplete)
What is the limit, and will it be increased? We have at least 2500 we want right off the bat.

[David]

Hard to know.
Like a lot of software companies we tend to under estimate the work involved in any project.

[robzilla]

When it does get released, there will be a free upgrade for people who purchased the software in the 6 months prior to the release date. For others who purchased a long time back there will be a discounted upgrade.

Would it be safe to assume that if I to order V6 today, I will get a free upgrade to V7 when it's released?

These latest developments seem very exciting, and frankly I would hate to miss out on them. I was reading about SFTP support earlier, but that wouldn't even be necessary if I can run the indexer on my Linux server with V7.

Great to see a site search engine being actively developed, I lost count of how many I've tried unsuccessfully.

[David]

No release date as yet. We are in the middle of a large code restructuring to split up the under interface and the engine. Once the engine is separate from the UI, then our plans are to port the engine to Mac and Linux. Then build a new user interface in PHP (in addition to keeping the interface that we have at the moment).

The end result should be that you can use Zoom on a Linux server without also needing a Windows desktop machine. After that we'll also be looking at doing a Mac desktop version.

Yes, we'll probably also need to have another look at where auto-complete can source its data from.

[Josiahjg]

auto complete and release date?

Regarding auto complete, this is a great feature! Can you add to the auto complete logo, or have it pull from META tags? For instance, we'd want all our item codes and item names in the list.

Is there a release date for V7 yet?

[Ray]

V7 Alpha 4 is now finally available:
http://www.wrensoft.com/zoom/v7.html

We have added SSL support so HTTPS is now supported (we had replaced the HTTP stack in V7, so that we no longer use the WinInet implementation).

We have also added support for Integrated Windows Authentication (NTLM).

- Fixed the PST indexing components which was not registering properly on a new machine.
- Added handling of .msg attachments (will index to indefinite depth)
- Fixed bug with corrupted URLs when indexing PST attachments
- Fixed bug with HTTP timeout waiting for over a minute
- Various fixes with comment filtering and tags
- Added support for both exiftool and ImageInfo plugins
- Fixed bug with spanish date formats
- Fixed bug with Japanese character
- Fixed bug with HTTP headers being indexed as part of the page
- Fixed handling of pages that did not report their file size via the HTTP header
- Fixed bugs in dictionary causing first variant to be skipped over.

And other various updates. Please let us know if you have any trouble with the new build.

[MergeThis]

As we provide our WebHelp to multiple customers (for installation on their own servers) will the Autocomplete feature work seamlessly?

That is, will each customer's experience be restricted to their users' search actions, since they will each be starting with fresh log and zdat files?

And what about upgrades (to the help along with our product)? Will we require our customers to backup those files and use them to overwrite the new ones?


Thx,
Leon

[David]

Tsui,

V7 is not yet released. We are still doing development. We expect to still be doing development for a couple more months at least.

When it does get released, there will be a free upgrade for people who purchased the software in the 6 months prior to the release date. For others who purchased a long time back there will be a discounted upgrade.