Kinook Software Forums

Go Back   Kinook Software Forums > Ultra Recall > [UR] General Discussion

Reply
 
Thread Tools Rate Thread Display Modes
  #16  
Old 04-30-2011, 03:42 AM
mlwang mlwang is online now
Registered User
 
Join Date: 04-27-2011
Posts: 34
slow launch culprit found (sort of)

Turned out reinstalling Windows didn't really solve the issue. Ultra Recall was quick to launch when Win 7 was fresh, but soon the slow launch I observed the last few days crept back. After many trials and errors, I've finally pinned down the cause, sort of.

It's a combination of intel's RST driver ver. 10 and Realtek's LAN driver ver. 7 that causes UR's launch delay. I've tried 3 different versions of RST, including the current official version (10.1.0.1008), a newer WHQL version, and the newest beta (the latter two from station-drivers.com). I've also tried two different versions of Realtek LAN driver, including the one on Asrock's web site (my motherboard is Asrock P67 Extreme 6), and the newest official release from Realtek's web site. The results were all the same. I couldn't try older versions of RST; they refused to be installed on a P67 board.

No wonder I saw the problem quickly after I started rebuilding the system; the drivers are among the first things installed with a new system.

Now, I can't pin the blame squarely on either driver, for installing one of them alone would be no problem. There's serious delay at UR launch only after both drivers have been installed.

For the moment, I've removed the RST driver since I would lose my network connection without the LAN driver. Windows 7's native AHCI driver seems to work ok, though it's generally recommended to go with the RST driver when using an intel SSD on an intel chipset-based motherboard.

I surely hope someone with similar hardware would test it out, thanks.
Reply With Quote
  #17  
Old 04-30-2011, 07:55 PM
mlwang mlwang is online now
Registered User
 
Join Date: 04-27-2011
Posts: 34
Quote:
Originally posted by kinook
There's probably an issue with the encoding Firefox uses when copying to the clipboard. [/B]
Firefox doesn't specify "lang=ja" as IE does. What Firefox does is correct, IMHO, for mixed-language web pages. Japanese wikipedia uses Japanese as the main language, of course, but it also often contains text in English, Chinese, and other languages.

Anyway, got it working by tweaking IE's font settings. Didn't realize Ultra Recall is so intricately related to IE.
Reply With Quote
  #18  
Old 04-30-2011, 08:06 PM
mlwang mlwang is online now
Registered User
 
Join Date: 04-27-2011
Posts: 34
Re: Re: newbie questions

Quote:
Originally posted by kinook
http://www.kinook.com/Forum/showthre...?threadid=4126 [/B]
Tried the custom Form method as described in the linked thread. It's not intuitive, but the instruction is detailed enough to follow step by step. Much appreciated.

Taking a look at the list of needs mentioned in my original post, however,

Quote:
1) I have to unhide other attributes if I want to see them, and coming back and forth is a bit cumbersome;

2) I can't click on the URL to launch it in the browser; and

3) when opening the document externally (in a browser), and you decide you want to see the source page, you have to come back to UR to click "Click here to open linked document". Wouldn't it be easier if one could just click on a link at the bottom to launch it without leaving the browser?
and it should be apparent that the 3rd need can't be satisfied by this method, and the second need is only partially helped. It would be much easier to be able to just click on a link to open the source url in a browser.

Anyway, thanks for the great trick. Please consider my request as a feature wish for future releases. Thanks.
Reply With Quote
  #19  
Old 04-30-2011, 09:36 PM
mlwang mlwang is online now
Registered User
 
Join Date: 04-27-2011
Posts: 34
Quote:
Originally posted by kinook
Quote:
Originally posted by mlwang
A couple of unicode-related oddities:

1. there seems to be issues with searching in Chinese/Japanese text? Some can be found straight up, some can only be found with an asterisk at the end, and still some others need to be sandwiched by a pair of asterisks. Is this a known issue?
No.
Could you please copy the following text into UR as a text item (I think web page item will do as well, but I tried it as a text item)?

Quote:
最前面加上一些與日文漢字不同的中文字,如「營、雜、舊」等等。And then some English text.

縄文時代

縄文時代(じょうもんじだい)は、年代でいうと今から約1万6,500年前(紀元前145世紀)から約3, 000年前(紀元前10世紀)、地質年代では更新世末期から完新世にかけてA列島で発展した時代であり、世 界史では中石器時代ないし新石器時代に相当する時代である。旧石器時代と縄文時代の違いは、土器の出現や竪 穴住居の普及、貝塚の形式などがあげられる。 草創期・早期・前期・中期・後期・晩期の6期に区分される。この頃の日本列島人は縄文式土器を作り、早期以 降定住化が進んで主に竪穴式住居に住んだ。弓矢を用いた狩猟、貝塚に見られる漁労、植物の採集などで生活を 営み、打製石器、磨製石器、骨角器などを用いた。

日本の歴史(にほんのれきし)、日本史(にほんし)とは、日本または日本列島における歴史、国史(Nati onal History)のこと。本項では日本の歴史を概観する。

各時代の詳細は、各時代区分項目(各節の冒頭のリンク先)を参照されたい。

前幾天看了報導才發現這次出宏碁報告的麥格理 "產業" 分析師竟然是我過去在宏碁的同事張家福
當年他還只是一個剛出社會歷練的台大畢業生,想不到10年後卻由他對宏碁發出重重的一擊
Now, with full text search enhancement enabled,
1) the term "產業" (which means "industry") can be found straight up,
2) the term "日本" (Japan) can be found with an asterisk appended at its tail, and
3) the term "日文" (Japanese the language) can be found only when sandwiched by a pair of asterisks (all searched without the quotation marks).

The 3rd scenario applies to most of the Chinese or Japanese terms in my tests.

With full text search enhancement disabled and the "Match whole words" option unchecked, both ""產業" & "日本" can be found straight up, while "日文" can only be found with ONE asterisk at the end.

Now, one related questions: why can't "Match whole words" be disabled temporarily (thus working without the full text search enhancement)? It isn't practical to re-compact a large database between searches.
Reply With Quote
  #20  
Old 05-02-2011, 09:57 AM
kinook kinook is online now
Administrator
 
Join Date: 03-06-2001
Location: Colorado
Posts: 5,750
For #1, the characters on both sides are word separators, for #2, only the character on the left side is a word separator, and for #3, the characters on both sides are not word separators, and * is required to match in the middle of words.
Reply With Quote
  #21  
Old 05-02-2011, 07:20 PM
mlwang mlwang is online now
Registered User
 
Join Date: 04-27-2011
Posts: 34
Quote:
Originally posted by kinook
For #1, the characters on both sides are word separators, for #2, only the character on the left side is a word separator, and for #3, the characters on both sides are not word separators, and * is required to match in the middle of words.
Is this something can you may address in a future release? Having to use asterisks all the time is cumbersome, but it's tolerable if there's hope that it's only a temporary workaround.

BTW, the word "告" in the sample text I gave above couldn't be found no matter how I tried (with or without asterisks) if full text search enhancement is disabled (and the "Match whole words" option unchecked), what's with that?
Reply With Quote
  #22  
Old 05-03-2011, 08:29 AM
kinook kinook is online now
Administrator
 
Join Date: 03-06-2001
Location: Colorado
Posts: 5,750
Quote:
Originally posted by mlwang
Is this something can you may address in a future release? Having to use asterisks all the time is cumbersome, but it's tolerable if there's hope that it's only a temporary workaround.
We will consider it.
Quote:

BTW, the word "告" in the sample text I gave above couldn't be found no matter how I tried (with or without asterisks) if full text search enhancement is disabled (and the "Match whole words" option unchecked), what's with that?
I'm not sure, but non-FTS support is only there for backward compatibility with older databases.
Reply With Quote
  #23  
Old 05-04-2011, 04:16 AM
mlwang mlwang is online now
Registered User
 
Join Date: 04-27-2011
Posts: 34
Quote:
Originally posted by kinook
We will consider it.
For your information (if you didn't know), indexing Chinese/Japanese words shouldn't be hard. Just index every character, for each of them is a word. While there are hundreds of thousands of characters, most people regularly use only around 3-5 thousands (for Chinese, and much less for Japanese). I believe a typical UR database has more than that number of English words indexed.

Anyway, thanks for taking it under consideration, and after I've devised a VB script myself to convert xml outputs into html pages with keywords and URLs appended at the end, I think I'm almost ready to take the plunge.

The only hurdle left is the relatively high price tag of the Pro version. May I ask (sheepishly) two more questions:

1. Is there any chance UR will be featured on BitsDuJour (or better, grant DonationCoder members a discount) in the near future? I searched and found it was featured on BitsDuJour a little more than a year ago, and wonder if I'd have the luck to see it there again in a few weeks. More than a hundred people (including me) signed up asking for the deal as I checked just now.

2. If the above request is out of question and I have to settle for the Standard version, is there a way to upgrade to the Pro version at a discount in the future? (Couldn't see such an option on the Order page.)

Thanks for your patience with my many questions and requests.
Reply With Quote
  #24  
Old 05-04-2011, 06:35 AM
kinook kinook is online now
Administrator
 
Join Date: 03-06-2001
Location: Colorado
Posts: 5,750
Quote:
Originally posted by mlwang
For your information (if you didn't know), indexing Chinese/Japanese words shouldn't be hard. Just index every character, for each of them is a word. While there are hundreds of thousands of characters, most people regularly use only around 3-5 thousands (for Chinese, and much less for Japanese). I believe a typical UR database has more than that number of English words indexed.
We use SQLite FTS3 for this. Either we're doing something wrong, or it doesn't index character by character for Chinese/Japanese text, but I believe that a match whole words option would resolve it.
Quote:
Anyway, thanks for taking it under consideration, and after I've devised a VB script myself to convert xml outputs into html pages with keywords and URLs appended at the end, I think I'm almost ready to take the plunge.
It would be great if you could post this at http://www.kinook.com/Forum/forumdisplay.php?forumid=28 -- thanks.
Quote:
The only hurdle left is the relatively high price tag of the Pro version. May I ask (sheepishly) two more questions:

1. Is there any chance UR will be featured on BitsDuJour (or better, grant DonationCoder members a discount) in the near future? I searched and found it was featured on BitsDuJour a little more than a year ago, and wonder if I'd have the luck to see it there again in a few weeks. More than a hundred people (including me) signed up asking for the deal as I checked just now.
It was last featured there on Jan. 10, 2011, and will probably not be back for several months. We'll consider offering a DC discount. If you are a student or educator or work for a non-profit organization, we can also offer you the educational discount (30%) at http://www.kinook.com/ordere.html.
Quote:
2. If the above request is out of question and I have to settle for the Standard version, is there a way to upgrade to the Pro version at a discount in the future? (Couldn't see such an option on the Order page.)

Thanks for your patience with my many questions and requests.
You can upgrade from Std to Pro in the same major version using the Pro Upgrade order option ($50).
Reply With Quote
  #25  
Old 05-04-2011, 09:04 AM
mlwang mlwang is online now
Registered User
 
Join Date: 04-27-2011
Posts: 34
Quote:
Originally posted by kinook
We use SQLite FTS3 for this. Either we're doing something wrong, or it doesn't index character by character for Chinese/Japanese text, but I believe that a match whole words option would resolve it.
With "Enhanced full-text Search" enabled, the "match whole words" option is always checked and grayed out (i.e., the setting can't be changed). Or am I doing anything wrong?

As to FTS3, I've absolutely no idea how it works. I'm not a programer; I can do only a little scripting. For now I simply pad all Chinese/Japanese search terms with two asterisks. I would be very grateful if you may turn your attention to this matter someday. Feel free to call on me to help testing when you do. I'm not a programer, but I'm pretty good at debugging. I've helped several programers making their products compatible with East Asian text over the years.

Quote:
Originally posted by kinook
It would be great if you could post this at http://www.kinook.com/Forum/forumdisplay.php?forumid=28 -- thanks.
I'd love to, except my script relies on a shareware editor (EmEditor) for its objects (mostly the selection object). I guess it can be modified to work with MS Word (using Word's selection object) but I haven't tried it.

In addition, the script is a quick hack put together in half an hour and tested only on a few test notes in my test database. I'm not sure if it'll work as expected for others. As an example, I found out during my test that UR sometimes stores the URL in a format like "![CDATA[http://...]". I've no idea what the "CDATA" notation is for, and tell my script to strip it. It might not be the wisest thing to do, and I've no idea what other special notations might go into that field.

Now, given the caveats above, do you still think it's suitable to be published in the Tips forum?

Quote:
Originally posted by kinook
It was last featured there on Jan. 10, 2011, and will probably not be back for several months. We'll consider offering a DC discount. If you are a student or educator or work for a non-profit organization, we can also offer you the educational discount (30%) at http://www.kinook.com/ordere.html.
Jan. this year? Too bad I didn't know it (nor did I have the time to try out UR at that time). But the educational discount is good enough for me (I'm a teacher) and I've put in my order and received my reg. info. just now. Thanks a lot for the discount.
Reply With Quote
  #26  
Old 05-04-2011, 09:57 AM
kinook kinook is online now
Administrator
 
Join Date: 03-06-2001
Location: Colorado
Posts: 5,750
Quote:
Originally posted by mlwang
With "Enhanced full-text Search" enabled, the "match whole words" option is always checked and grayed out (i.e., the setting can't be changed). Or am I doing anything wrong?
No, that option isn't currently available with FTS enabled, and you must use *'s to achieve it. But we'll probably address this problem by enabling that option and adding *'s behind the scenes.
Quote:
I'd love to, except my script relies on a shareware editor (EmEditor) for its objects (mostly the selection object). I guess it can be modified to work with MS Word (using Word's selection object) but I haven't tried it.

In addition, the script is a quick hack put together in half an hour and tested only on a few test notes in my test database. I'm not sure if it'll work as expected for others. As an example, I found out during my test that UR sometimes stores the URL in a format like "![CDATA[http://...]". I've no idea what the "CDATA" notation is for, and tell my script to strip it. It might not be the wisest thing to do, and I've no idea what other special notations might go into that field.

Now, given the caveats above, do you still think it's suitable to be published in the Tips forum?
I could be helpful as-is. The CDATA sections are an XML construct (see http://en.wikipedia.org/wiki/CDATA) that will be used for text that contains markup characters. If you use an XML parser to work with the document, it should be handled transparently.
Reply With Quote
  #27  
Old 05-05-2011, 11:54 PM
mlwang mlwang is online now
Registered User
 
Join Date: 04-27-2011
Posts: 34
Quote:
Originally posted by kinook
But we'll probably address this problem by enabling that option and adding *'s behind the scenes.
Thanks in advance.

Quote:
Originally posted by kinook
I could be helpful as-is. The CDATA sections are an XML construct (see http://en.wikipedia.org/wiki/CDATA) that will be used for text that contains markup characters. If you use an XML parser to work with the document, it should be handled transparently.
Thanks for the pointer, so I guess it's ok to remove it for my purpose (injecting into .html files).

Just made some modifications and add a few lines of notes and comments in the script so hopefully it'll be more comprehensible to others. It's been posted at http://www.kinook.com/Forum/showthre...&threadid=4846
Reply With Quote
Reply

Thread Tools
Display Modes Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



All times are GMT -5. The time now is 11:02 PM.


Copyright 1999-2019 Kinook Software, Inc.