Improved search results

Forum for the PDF-XChange Editor - Free and Licensed Versions

Moderators: TrackerSupp-Daniel, Tracker Support, Paul - Tracker Supp, Vasyl-Tracker Dev Team, Chris - Tracker Supp, Sean - Tracker, Ivan - Tracker Software, Tracker Supp-Stefan

Post Reply
Timur Born
User
Posts: 874
Joined: Tue Jun 26, 2012 1:50 pm

Improved search results

Post by Timur Born »

10 years ago was the first time I suggested to implement search results sorting/filtering. We are 5 days short of a proper anniversary, but a more current discussion is a welcome reason to dig this up again:
Timur Born wrote: Tue Jun 26, 2012 2:08 pm Hello everyone!

For me PDF-XChange Viewer beats about every other Windows based PDF browser on the market when it comes to rendering and features! Unfortunately its Search functionality is just as basic (out of the last century) as everything else on the market. Believe it or not, but main reason for me to use OS X from time to time is for getting proper Search functionality for large PDF books.

Basically nearly every Windows PDF viewer only offers to go chronologically back and forward in Search results and even only lists results in chronological order. That's as if we would search the Internet via back and forward one item after the other and makes searching a very time consuming effort.

On OS X the built in Preview PDF viewer and some third party alternatives offer search results that are ordered by relevance (or at least number of hits per page), include thumbnails of relevant pages and colored markings of all found words on a page.

Please consider adding a real and modern Search functionality into PDF-XChange Viewer so that I don't need to fire up OS X just for better PDF handling.

Thanks and regards!
TrackerSupp-Daniel wrote: Mon Jun 20, 2022 3:52 pm Hello, Timur Born

I can forward these points for consideration to our Development team, but I would advise against getting your hopes up too high here. Offering a chronological list is much more straightforward and predictable than having hinting and relevant searches ongoing to muddle the output. It may seem like it would be better, but the criteria for such "relevance sorting" is FAR more complex than it would appear to be, and would be just as likely to appear almost random, as it would to help in some cases.
It is far more liable to annoy and confuse a large portion of our users who are expecting the simple, straightforward chronological list, if they updated and encountered what you are suggesting.

It may be something that is reconsidered and offered in the far future, but it is not likely to be something that we see for many years yet.

Kind regards,
I disagree strongly, especially when we are talking about additional options that leave the current system intact. Sorting by relevance would take bookmarks and headlines into account and then maybe number of hits.

Most of the times I search for a term the most relevant result is the one that points to a corresponding bookmark which in turn points to the corresponding page that includes corresponding headline. Next in relevance are those results with no corresponding bookmark, but where a chapter's/paragraph's headline already contains the search term or maybe even consists solely of the search term (in my documents even written ALL in capitals). Last, but not least the chances to find relevant content is higher on pages/paragraphs with multiple/more hits of the search term.

Currently Editor's search results take *none* of these into account and does not even allow to use bookmarks in search results to directly go to their corresponding page. I do not want to be forced to search PDF files chronologically anymore than I would want to be forced to search the internet chronologically. And if the first search result is a bookmark (with a high chance of pointing to what I am searching for) then at least allow me to CTRL-click or double-click on that bookmark search result to navigate directly to its corresponding page instead of forcing me to do first go through the bookmark panel.

We are living in the age of big data, with hundreds and thousands of PDF pages being searched for relevant results. PDF software's job is to make these tasks easier and I do not have to fire up MacOS just for better PDF search results. Nowadays even less so than 10 years ago, because I usually avoid MacOS and because its own search results did not improve either.
User avatar
TrackerSupp-Daniel
Site Admin
Posts: 8436
Joined: Wed Jan 03, 2018 6:52 pm

Re: Improved search results

Post by TrackerSupp-Daniel »

Hello, Timur Born

We have gone down roads with these types of discussions before. The simple fact is that searching by relevance, is vastly more complex than you are making it out to be, and with the manpower we have, it would not be something that we can just put together in a few weeks time. Even if we did do so, the results would not be perfect, and would take years more of trial and error, and reactive improvement as bugs are reported. The reason that relevance based search works well in web browsers is because of manually defined "tags" that are used to increase accuracy, with a PDF search function, we would need to programmatically detect and apply these before you can begin searching, or the results would be ineffective.

As an example, If I am searching for "orange" on the internet, what will appear first, large blank images of the color orange, or the Wikipedia article about Oranges, the Fruit? Relevance sorting is done by guessing context and then looking through the results. Yes, in a PDF about Fruit I could search for orange and get results across the board about oranges, but looking at the bookmarks labelled as "thick skinned fruit" and "thin skinned fruit" will not help in any capacity, and trying to use those for sorting, before showing the remaining results would result would simply mean that the content is out of order, not necessarily any better presented. The "thick skinned" bookmark does not have any "tags" within it saying that it encompasses "oranges, apricots, bananas, etc".

I cannot deny that Yes, you are correct, in many cases even rudimentary relevance sorting could be an improvement, but in just as many cases, it would be worse, and that is before we factor in the sheer time and effort that would be required to get it right. We do not currently intend to make this change. It is worth considering that we have millions of users, not all of them use the software the same way you do, and any changes on one side of the fence, often will have consequences on the other side of it.

If something like this is offered in the future, then you can rejoice; for now, this feature request has been expressly rejected by the development team, and only exists as a placeholder for future reconsideration.

Kind regards,
Dan McIntyre - Support Technician
Tracker Software Products (Canada) LTD

+++++++++++++++++++++++++++++++++++
Our Web site domain and email address has changed as of 26/10/2023.
https://www.pdf-xchange.com
Support@pdf-xchange.com
Timur Born
User
Posts: 874
Joined: Tue Jun 26, 2012 1:50 pm

Re: Improved search results

Post by Timur Born »

As mentioned before, even the simple option to directly jump to bookmarks destinations from search results would help minimize eye-movement, mouse-movement and clicking. And sorting by number of hits per page/paragraph could also be useful.

Here is a search for the word "prone":
Image
47 hits with "words on same page" proximity search. The second hit points to a bookmark called "Prone".

Clicking on the search result scrolls the bookmark panel to the corresponding bookmark, leading my eye and mouse-pointer from the upper right to the lower left.
Image

Once I click on the bookmark on the left it opens a page with a paragraph titled "Prone" (sole word in its own line). This is search result hit number 41 out of 47, with a total of 6 occurrences of the word "prone" on the same page. A double-click (or ctrl-click) on the bookmark search result would have gotten me there easier.

Here is an example of a word that does not match a bookmark. But the best hit is the one where the search matches is the (capital) headline and a second hit on the same page in the text-body right below the headline. Both strong indicators for a good hit compared to the higher listed search results.

Image

Furthermore, some tablet based PDF viewers can filter thumbnails by search results, as can MacOS Preview, which is also helpful when the document reveals content via page formatting.

All PDF Viewers/Editors out there concentrate on annotations and whatnot, but those of us users who mostly want to read and search PDF files are unfortunately left in the digital middle-ages.
User avatar
TrackerSupp-Daniel
Site Admin
Posts: 8436
Joined: Wed Jan 03, 2018 6:52 pm

Re: Improved search results

Post by TrackerSupp-Daniel »

Hello, Timur Born

I am sorry, but I cannot, at this time, give you the news you are hoping to hear. This feature is not planned to be implemented due to the complex implementation requirements and large time investment that is required. It has (very recently) been rejected by the Dev team and will not be reconsidered for quite some time.

Kind regards,
Dan McIntyre - Support Technician
Tracker Software Products (Canada) LTD

+++++++++++++++++++++++++++++++++++
Our Web site domain and email address has changed as of 26/10/2023.
https://www.pdf-xchange.com
Support@pdf-xchange.com
Timur Born
User
Posts: 874
Joined: Tue Jun 26, 2012 1:50 pm

Re: Improved search results

Post by Timur Born »

I will come back to this after another 10 years then.
User avatar
Paul - Tracker Supp
Site Admin
Posts: 6829
Joined: Wed Mar 25, 2009 10:37 pm
Location: Chemainus, Canada
Contact:

Improved search results

Post by Paul - Tracker Supp »

:(
Best regards

Paul O'Rorke
Tracker Support North America
http://www.tracker-software.com
Timur Born
User
Posts: 874
Joined: Tue Jun 26, 2012 1:50 pm

Re: Improved search results

Post by Timur Born »

:P
User avatar
Paul - Tracker Supp
Site Admin
Posts: 6829
Joined: Wed Mar 25, 2009 10:37 pm
Location: Chemainus, Canada
Contact:

Re: Improved search results

Post by Paul - Tracker Supp »

:lol:
Best regards

Paul O'Rorke
Tracker Support North America
http://www.tracker-software.com
Timur Born
User
Posts: 874
Joined: Tue Jun 26, 2012 1:50 pm

Re: Improved search results

Post by Timur Born »

1.5 years later I still have to move eyes and mouse from upper right to lower left in order to get to the destination of a bookmark found in search. Also no other improvements to searching large PDF files more conveniently.

My maintenance plan ends in 3 days and I will not renew it for the time being.
User avatar
TrackerSupp-Daniel
Site Admin
Posts: 8436
Joined: Wed Jan 03, 2018 6:52 pm

Re: Improved search results

Post by TrackerSupp-Daniel »

Hello, Timur Born

If you simply wish to reduce the overall eye/mouse movement, you could affix the bookmarks pane to the right side of the screen, keeping it close to the search pane to minimize that movement. There is no rule that says it needs to reside on the far left of the application. But I am afraid the stance on this remain the same as before. Search is already overcomplicated and overloaded with options, every change we make there needs to be carefully considered and this one has been expressly rejected in the past.

Kind regards,
Dan McIntyre - Support Technician
Tracker Software Products (Canada) LTD

+++++++++++++++++++++++++++++++++++
Our Web site domain and email address has changed as of 26/10/2023.
https://www.pdf-xchange.com
Support@pdf-xchange.com
Timur Born
User
Posts: 874
Joined: Tue Jun 26, 2012 1:50 pm

Re: Improved search results

Post by Timur Born »

I am not convinced that allowing a double-click (or modifier+click) on a thumbnail search result would add to the complexity of the search results pane. It surely would beat this in convenience:
image.png
User avatar
Tracker Supp-Stefan
Site Admin
Posts: 17820
Joined: Mon Jan 12, 2009 8:07 am
Location: London
Contact:

Re: Improved search results

Post by Tracker Supp-Stefan »

Hello Timur Born,

I would ask our devs to review your request, however I can't really make any promises at this time!
Please note that the search can find the name of the bookmark, but the bookmark itself might execute more than just the action to take you to a specific page in the document. So really I am not too sure if such a "execute bookmark by clicking it's name in the search pane" should be offered. I know that the majority of cases would be where the bookmarks only take you to other pages in the same file, but still - there can be exceptions, and we need to care for all cases equally!

Kind regards,
Stefan
Post Reply