New release of AntConc (4.4.2)

[Laurence Anthony]

Dear All,

I'm happy to announce the release of AntConc (Version 4.4.2). This version is available for Windows, Windows Portable, MacOS (Silicon), and Linux (tested on Linux Mint).

This release introduces an automatic updater for older corpus databases that were not compatible with the 4.4.x version, so they can be now used seemlessly.  You should also find that the startup time is now much faster. I've also fixed various bugs and optimization issues, so you should find that the new version works much more smoothly than 4.4.1.

You can get the new version here:

https://www.laurenceanthony.net/software/antconc/

I really hope you like the new version!

Best regards,

Laurence.

[Rudy Loock]

Dear Laurence,

I hope you are doing well. Thank you for this new release of AntConc. I am encountering 2 problems, though, one of which is quite imoprtant. First, it was impossible to install it on my desktop PC (Windows 11) without unintsalling the previous version (this has never happened before). Even then, I got several messages telling me that the app couldn't find the right directory. Now that I have launched it several times it seems to work.

But what worries me the most is that there seems to be a problem with the definition of token. For instance, when I search for a term (e.g. hantavirus) AntConc dismisses all occurrences where the word is followed by a sign of punctuation: "hantavirus." or "hantavirus," are not retrieved for instance. I compared with version 4.0.0 and this issue does not happen (note that I didn't change the token definition settings when compiling the corpus). The only way I can make sure to retrieve all occurrences is by searching for "hantavirus*". Again, this did not happen with previous versions of AntConc and is quite counter-intuitive for students. Do you have a solution for this?

Best wishes,

Rudy

[Laurence Anthony]

Hi Rudy,

Sorry for the super late reply. I was away in Australia over the summer and just got back home after more work and conferences here in Japan.

Let me address a the issues you talk about:

>First, it was impossible to install it on my desktop PC (Windows 11) without unintsalling the previous version (this has never happened before). Even then, I got several messages telling me that the app couldn't find the right directory. Now that I have launched it several times it seems to work.

What version did you have on your computer before the install, and what version were you trying to install over it? Windows is actually really quite messy when it comes to installations, and my guess is that some old dependency was stuck in the folder and was preventing a smooth update. I've had the issue appear here occasionally, but let me know if this starts becoming repetitive. Also, it's weird that the app reported that it couldn't find the correct directory. It suggests that there might have been a more general issue with your install. You're the first to report this, so let's monitor this over the next few weeks as I make further updates.

>But what worries me the most is that there seems to be a problem with the definition of token. For instance, when I search for a term (e.g. hantavirus) AntConc dismisses all occurrences where the word is followed by a sign of punctuation: "hantavirus." or "hantavirus," are not retrieved for instance.

Yes, this is very problematic. Can you confirm what version you are using, and also send me a simple 2-3 sentence file that shows this problem. If I can replicate it here, it should be trivial to solve. Note that I can't replicate the problem here. As you can see below, "cat" is showing correctly as three hits:

If your problem is really there, it is very urgent to fix it, so I'll make this my highest priority.

Regards,

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

--
You received this message because you are subscribed to the Google Groups "AntConc-Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to antconc+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/antconc/d9b7d64e-3663-431c-ac2b-d0de326896een%40googlegroups.com.

[Rudy Loock]

Dear Laurence,

Thank you for your reply.

As for the first question, I had version 4.4.0 when I tried to install 4.4.2. Like I said, I did manage to install it but there were issues that required the uninstallation of version 4.4.0 before and this was new.

As for the second issue, here is the data. Based on the same documents, I compiled the same corpus (French) on versions 4.4.0 (on my PC laptop) and 4.4.2 (on a PC desktop). First surprise: the number of tokens is very different: 10,329 for 4.4.0 and 9,582 for 4.4.2. When I search for a term (hantavirus), I get 139 hits with version 4.4.0 but only 81 with 4.4.2. From what I can see, it is the hits with a punctuation mark that are missing. If I try hantavirus*, it partly solves the problem since I collect 99 hits (but still not 139). I tried with other search terms and the problem is the same.  The same issue shows with English data (147 hits for hantavirus with version 4.4.0 and 127 with 4.4.2). I hope this helps clarify things (note that when I compiled the two corpora, I used the by default settings for token definition).

 

Best wishes,

Rudy

[Laurence Anthony]

Hi Rudy,

Did you forget to attach the data?

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

To view this discussion visit https://groups.google.com/d/msgid/antconc/299c8610-1190-4d5f-afd1-d145f0b8fc83n%40googlegroups.com.

[Rudy Loock]

Hi Laurence. No, I did not provide any attachment. What would you like? Screenshots with the results or the .db files for the corpus?

[Laurence Anthony]

Hi Rudy,

As I wrote earlier, could you "send me a simple 2-3 sentence file that shows this problem."

You can also email your .db file and the search query you used. But, shrinking the problem to the smallest possible file is always the best first step.

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

To view this discussion visit https://groups.google.com/d/msgid/antconc/77cda710-46cc-4e85-a7e4-942302b9fa7bn%40googlegroups.com.

[Rudy Loock]

OK, here is something simple that shows the problem easily I think. Based on the 2 .txt files documents in attachment, if I search for influenza with AntConc 4.4.0, I get 2 results, which are both followed by a punctuation mark (inflenza, and influenza.). If I search with version 4.4.2, I get No hits found.

[Laurence Anthony]

Hi Rudy,

I just tested your files here. Everything seems fine to me.

I wonder if you have some setting that is causing this. Can you try restoring the default settings from the File Menu and repeating the corpus construction and search?

Regards,

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

To view this discussion visit https://groups.google.com/d/msgid/antconc/afd5a192-be5e-455f-a5a6-b99be7cbee10n%40googlegroups.com.

[Rudy Loock]

Dear Laurence,

I did as you suggested and it seems to have solved the problem! I am really wondering which setting I may have changed to create this problem (I really don't remember changing anything to be honest, and this was the very first corpus I created with version 4.4.2, and I did compile it several times in case something went wrong). So I really don't understand how this could have happened; could it be a consequence of the installation issue I mentioned?

Note that I still have a slight difference between versions 4.4.0 and 4.4.2 (I did restore default settings and compile the corpus again for both), e.g. for the same search term, 139 hits for 4.4.0 (this has not changed) vs. 144 hits for 4.4.2.

Anyway, thank you so much for your help; I hadn't thought of restoring default settings!

Best wishes,

Rudy

[Laurence Anthony]

Hi Rudy,

It's great that the major issue seems to be resolved. There are many settings that you might have changed, so it's difficult to know exactly what happened. 

But, I'm still a bit worried about the following:

139 hits for 4.4.0 (this has not changed) vs. 144 hits for 4.4.2.

The number of hits should be identical. 

Again, if you can send me the exact data and the exact search, I can check this. One great way to resolve an issue like this is to save the .db file immediately after the search as all the rows and even the search settings are stored in the .db. So, if you have the .db for 4.4.0 and 4.4.2 after the same search, we can explore the differences very precisely.

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

To view this discussion visit https://groups.google.com/d/msgid/antconc/0cf695a4-f7e1-471b-9770-72199f361fd7n%40googlegroups.com.

[Rudy Loock]

Here are the 2 .db files for versions 440 and 442. The search term used is hantavirus.

[Laurence Anthony]

Hi Rudy,

I just checked the two corpora and they are not the same.

For 442, the corpus size is 9771 (tokens) and 2004 (types).

For 440, the corpus size is 10329 (tokens) and 2023 (types).

I've also found the probable cause. The PDF to text converter was improved between 440 and 442, so some texts and now being better converted, producing full words instead of messy text fragments. This is actually a good case for why is often better to create your corpus, clean it, and then load it into a corpus analysis tool like AntConc. My guess is that if you use say AntFileConverter to convert the corpus first, you'll get the same results in both 440 and 442.

Have a look for "diagnostiquer". It doesn't appear in 440 because of the pdf conversion issue, but it appears properly in 442.

I hope that helps!

Laurence.

###############################################################

Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

To view this discussion visit https://groups.google.com/d/msgid/antconc/777ecbe0-ccf4-478c-8b4b-7331bcfa8809n%40googlegroups.com.

[Rudy Loock]

Thanks a lot Laurence, I hadn't thought of that. The 2 corpora were built within AntConc directly from the same documents without any conversion first. It is interesting to know that this can have consequences on the number of hits.
Thanks a lot for spending so much time on this issue!

Best wishes,

Rudy

[Laurence Anthony]

You're very welcome!

I think this highlights the importance of version numbering! But, it also means that the 440 results are not 'wrong', but the 442 results are more accurate.

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

To view this discussion visit https://groups.google.com/d/msgid/antconc/00b34a91-8dff-4e4a-9cb0-af3388825ba0n%40googlegroups.com.

[Duyen Ha My]

Hi Sir, 

I'm currently using the version 4.4.2. However, the duration for searching the KWIC to File View is extending significantly, from nearly 1s to 15s now. I've downloaded the old version 4.2.0 but this situation has not improved at all. Compared with the oldest version Ive used, the duration for searching from KWIC to File view (Order by freq) only takes 1s to process successfully. 

 I really appreciate your response 

Best regards, 
Louisa

Vào lúc 19:53:22 UTC+7 ngày Thứ Hai, 20 tháng 7, 2026, Laurence Anthony đã viết:

[Laurence Anthony]

Hi Duyen,

Thanks for the feedback. When I moved from the AntConc 3x series to 4x series, I had to change how the file view was generated. For short texts of a few thousands words, the file view should be as quick as before. In recent releases, I also improved the performance so that a whole book would render very quickly. As an example, open the demo corpus and view the Alice in Wonderland book. It should open in a fraction of a second. What size of file are you trying to view? Can you tell me how many words it contains?

Laurence.

###############################################################

Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

--

You received this message because you are subscribed to the Google Groups "AntConc-Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to antconc+u...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/antconc/094f03b9-1092-4638-9c69-9334d4de69fen%40googlegroups.com.