Hook version 1.5 (3331) is now available

Thanks for all the feedback, Rob. We’ll pull out the encryption and fix the issues.

I think it might be worth scrutinizing exactly what happens if there is a late/subsequent reactivation, and exactly how the license gets lost.

The old database file may have been there before I discovered that the license had been lost. It was only after receiving the new activation and clicking that that I realised that that the data was nowhere to be seen.

Thanks, Rob. We’ve run yet further tests here and confirmed that the database is not deleted (at least in our tests), and that the upgrade to 1.5 format works. Our dev team will verify if there’s an interaction going on with license activation failure (that would explain why we haven’t received other mentions of database issues; but we have had one other mention of license issue. It may be that there was a network failure during license activation).

Thanks for your patience with this.

It’s quite coincidental that you should point to that article , because I’ve been resisting for several days publishing an article on https://CogZest.com about Covid policies and beliefs being based on so called “evidence-based” models (and discussing same with an AI researcher/author colleague). I agree there are reasons to criticize evidence based policies. ( However, models are required. Researchers put out a lot of models. Policy makers need to carefully loook at the assumptions, and not naively latch onto a single model. Beyond stats we need good mechanistic models; and vice versa. ) My article was going to (and would if I take the time to write it), draw a contrast between Humean and Kantian approaches to knowledge. Both are required, but Kant tends to get ignored. If you look at https://CogZest.com you will see several references to Kant and Popper. Goes back to my undergraduate days when I wrote papers on causal reasoning (which was to me my phd topic at Sussex, Sloman my supervisor is an expert on that, but I latched onto his Cognition and Affect project instead). The key concepts are mechanism and reasoning, with an emphasis on mechanistic reasoning. You’ll see criticisms on CogZest of “cottage industries” in science, e.g., in contrast to Claude Lamontagne’s amazing AI mechanistic models of vision. (To give an example re covid, there are strong mechanistic reasoning why wearing masks would help prevent transmission. But also some more direct empirical evidence).

So we are actually on the same page. My Cognitive Productivity books and approach are more Kantian than Humean. Perhaps behind your many criticisms of the CogSci Apps brand is a belief that I am Humean/ inductivist. I’m relatively more Kantian and Popperian.

If you look at chapters 7, 13 and 14 of Cognitive Productivity, you will see strong arguments in favor of productive practice apps. This is another type of app that does not currently exist (though flashcard apps can be shoe-horned into productive practice workflows). We actually started devising such an app. So we have three app concepts that come out of an integrative design-oriented research programme. This is not to say that there is some kind of a direct line of inference between science and apps. I’ve pointed in previous posts to Donald Schön’s concept of reflective practice, which is also discussed in the first Cognitive Productivity book. He has a very strong argument that aligns with what might be your view and definitely is mine that between science and engineering a fair amount of “filling in the gaps” happens.

Another point is that the distinction between science and engineering are not that clear cut. Nor is the distinction between math and science. Chapter 2 of The Computer Revolution in Philosophy (by my phd thesis supervisor) is a great read on that subject.

As I mentioned before cognitive science does not tell the developer where to put the pixels or widgets. But it does provide a large body of knowledge with which to think about problems and their solutions. I think there is something to be said for the fact that out of an integrative design-oriented R&D programme, three highly original apps arose ( mySleepButton, Hook, and an unfinished productive practice app). One day if I can I would like to write about that. I’m happy that there is at least one person (yourself) who seems to be interested enough in the approach to challenge it.

But again, I don’t think there’s much of a point in me writing pages and pages about this on the CogSci Apps or Hook websites at this point.

1 Like

So far two reports of license issues with 1.5; and one report of a database opening issue. I’ve had reports of successful upgrades. We’re still investigating the DB / license issue.

There’s a bug in “has hooked bookmarks for” which we’ve now fixed internally.

Kantian and Popperian

Not sure that easier creation and finding of links is likely to be much helped by a theory of knowledge :wink:

That seems like a distraction, at best.

Focus on better design and execution is what it takes.

You mentioned that the Hook 1.4 database, hook.sqlite, is not in ~/Library/Application\ Support/com.cogsciapps.hook/. That would explain why your 1.5 Hook instance has no data. The reason is that Hook 1.5 imports the Hook 1.4 database into hook_en.sqlite. So if the source is empty of course the target will be empty.

Now the question is why would the hook.sqlite (1.4) DB be empty ? Hook has no code that would delete that database. Only a user action can do that. Is it possible you have some script, add on or other process that might have deleted it that file or manipulated the ~/Library/Application\ Support/com.cogsciapps.hook/ in some way?

Could you check whether the aliases folder in that folder has files? It should have plenty of files in it. Otherwise, my only explanation is that there is a process or user action that deleted it (not Hook).

We’ve asked other users and they have reported no issues with the Hook library files (the 1.4 or new 1.5 one).

I don’t yet know why you would have been prompted for a library file. Someone else reported a license issue, but not lost data.

That was a response to you mentioning The Guardian article , above in this topic, in relation to CogSci Apps (which coincidentally deals with a topic I’ve been mulling over, i.e., the epistemology of covid personal and gvt responses). Tim Cooke: “I spend almost all of my time on people, strategy and execution. And I think that most everything else falls from those.” Yes too to requirements and design :slight_smile: .

Getting back to the issue you reported: Looking at your screenshot, we see the time stamp of the 1.1 file is today. The 1.1 tag file was created before 1.5. This suggests that in that folder of yours, Hook created everything today, meaning that you started from a fresh installation of Hook. That would explain why you have no data there. Your alias folder being empty would confirm that. At this point, it does not look like an issue with Hook.

No. It wasn’t a fresh installation when I clicked the new activation link. It seems to have become one at that point.

I’m not using any scripts which reference that folder.

Whether it is a Paddle issue or a CogSci issue I have no idea – the mechanism which loses the license should be some kind of clue.

The only thing that I wonder about on the system at this end is whether the Little Snitch firewall blocked any component message of the re-activation process. As I reported:

its as if the original folder had simply been completely overwritten

If that wasn’t by the upgrade, which left me (and, I hear, at least one other) with an inactivated license, then it sounds like the reactivation.

Thanks for letting us know, Rob. I don’t see a detailed mechanism on our side right now, but we will persevere.

Could you please check the date added, created and modified of your ~/Library/Preferences/com.cogsciapps.hook.plist file?

You may have to explain to me what you have in mind with the distinction between ‘added’ and created, but I wonder if the preferences file is getting eternally created afresh ?

The upgrade was yesterday morning (Europe) (see the post times), and the created and updated stamps appear to be simultaneous – just after midnight last night:

and now, having experimentally launched and re-closed Hook, Date Modified and Date Created are both 11:40

Ah – Date Added – found it, though I’m not sure how it’s defined.

In any case, seems to be the same as Created And Modified - all three appear to be reset on each use of Hook.

All 12.39 in this case:

and if I launch and close Hook again, now all 13:44 today.

(macos Mojave 10.14.6)

(And no plists in that folder have any differences between their Modified, Created, and Added dates)

Thanks for the extra info, Rob.

  • just to be sure, when you encountered the missing data issue, you had upgraded from 1.4 to 1.5, not 1.3 to 1.4, correct? The reason we ask is that the 1.1 tag file was introduced in Hook 1.4. So, if you had been running Hook 1.4 , given the 1.1 tag is present this suggests that the entire Hook Library folder (~/Library/Application\ Support/com.cogsciapps.hook/) was created at 1.5 launch time.
  • could you please also let us know the date created and date added of ~/Library/Application\ Support/com.cogsciapps.hook/ ?
  • Per above, could you also please verify whether your ~/Library/Application\ Support/com.cogsciapps.hook/aliases/ folder contains any entry created before your run of Hook 1.5? If the Hook Application Support folder is new, then it should not contain such entries.

The license re-activation error you experienced is just a coincidence. LittleSnitch, a network problem, or the license activation server could cause the license message you saw. We have exercised that part of the code countless times in last two years, as have our users. We reexamined the code and ran more tests confirming this. There’s no causal path in Hook between a re-activation failure and reinitializing Hook’s library folder.

When Hook starts, if its Applications Support folder is not present, of course it will create it. However, once that folder exists, Hook will never recreate it or reinitialize it. There’s no code in Hook to do that. Even if your macOS instance gave Hook a false reading that the folder does not exist when in fact it does (extremely unlikely), Hook could not recreate or init that folder. (We’ve examined and tested for this).

  • a telling test would be to revert to Hook 1.4 (launching it once with the prior state) and then quit and launch Hook 1.5. It should convert the database and run smoothly.
  • it may be worth checking whether any other Library or other folder is missing from that user’s account. I’m sure you’re extremely well aware of diagnostic methods, but if someone were to experience issues reading and writing from folders, I might normally recommend Apple’s support pages including Resolve issues caused by changing the permissions of items in your home folder - Apple Support.

It will work on any system that has Hook (even the free, Lite, version of Hook) and has a file with the same name. One of the social benefits of Hook is that you can share links to files that are revealed in their Finder. They are relative to the terminal node (the “n=” field), also considering some of the path upward from there, preferring the best match.

Some of the use cases for this are sending links to shared files or folders that are in a checkout of a version control system (Git, SVN, whatever), or a locally synced cloud share (e.g., Dropbox).

This contrasts with sending http:// links to shared Dropbox files, which take you to the web, which is slow.

I often send hook://file links to colleagues and family members. It saves them having to search for the item. It means they are more likely to act on messages about files.

So in short the hook://file link I shared with you should work. Just copy and paste it in Spotlight.

CogSci Apps needs to do a better job of marketing the many features of Hook :slight_smile: . Presumably, as people start sharing hook://file links, the number of Hook users will increase.

This doesn’t work for me. Did you mean Spotlight? Pasting in Safari does a Google search and posting in spotlight just brings up the search history of when I tried it in Safari.

Pasting links into Spotlight per How to Open Website URLs from Anywhere on Mac with Spotlight.

However, now in my test (macOS 10.14.6 latest) Spotlight won’t serve web URLs, OmniFocus or Drafts either.

Anyway, as for the original hook://file link I posted way up there, if you paste it in Alfred or LaunchBar it does work. It should work in any software that honors app links (e.g., omnifocus:// , drafts:// or hook://). E.g., BBEdit has a ‘openURL’ function.

The reason I suggested using a tool like that is that Discourse neutralizes app links. I don’t know a way around that. So when I paste app links here, I escape them in ticks ``.

This is all working for me now — I didn’t know you can open URLs with Spotlight but have found out that 1) you need to include the protocol for it to work with web URLs 2) you have to wait for the results to load, simply hitting enter doesn’t resolve immediately as expected.

The issue I was having with the URL above is I was copying this:
Screenshot 2020-06-28 at 16.52.47
which had hidden debris at the end:
(hook://file/LBUnTs2tn?p=TGlicmFyeS9BcHBsaWNhdGlvbiBTdXBwb3J0&n=com.cogsciapps.hook) folder ( ~/Library/Application\ Support/com.cogsciapps.hook/).

1 Like