I think that sounds like a swell idea personally.
1 Like
I think that would be a good solution for a few databases, but if you have 10, 20, 30 or more (like with a business having a large client base), I think it would be less than ideal.
Just curious - does each client get a separate database? Why?
That is what I do. I have a lot of my set ups on a Keyboard Maestro palette and the native hotkeys on ‘Workspaces’. I have over ten at present. I have smart groups set up to show via hotkeys too. If, as it is for me, DEVONthink is really my main app it is worth setting those up. The advantage of palettes in Keyboard Maestro is that you can see what you need and don’t need to memorize or call up lists of hotkeys. Consequently I only have the sidebar open in the Global Inbox.
I am on DEVONthink 2 still but preparing. I only started using 3 pane view recently too funny enough, after three years of using the app! I loved it but it was very counter intuitive when I first started and I never used it.
1 Like
I don’t have that many clients currently, but if I were a lawyer or someone with a lot of clients, it would make sense to keep each of them in their own container, or database. This limits the cross-contamination of information, and as in the example above, helps prevent people from seeing information that they (ethically and perhaps legally) should not be seeing.
Even with only had a handful of active clients at a time, I’d be constantly managing which hot keys opened up which active client. I can easily see that becoming quite the time-consuming mess.
Not to beat a dead horse, but DT already had a great solution for this and it was removed. I’m confident the DT dev team will be able to figure out how to replace a lot of the functionality the 3PV had, but we’ll just have to wait and see.
You could use this script I wrote. It shows a list of all workspaces, then opens the one you select.
1 Like
Interesting - do others do that too?
What sort of cross-contamination of information are you concerned with?
Even in medicine, where rules about access to information are extremely strict, the standard is to keep all records in one patient or “client” database.
As for limiting access, unless you have a different password for each client the problem of unethical review of data still exists.
It seems to me there are also lots of housekeeping requirements such as backing up data - and verifying the backups - that would become impractical with such a large number of databases.
And it also seems to me that such a system is inherently non-scalable. What happens when the business with 30 clients grows to 100? 200? Or more?
I don’t do this (separate Db for each client), as I don’t currently have a need for it, but @evank made the point above about keeping his clients in separate databases, and why, and I was simply commenting, as it makes sense to me. Having dealt with clients in the past, I can definitely see the advantage of keeping them in a separate Db, where all of their email correspondence, attachments, notes, etc. is kept together.
Actually, I would think having each client in its own container (database) makes it very easy to archive when no longer needed, as well as restoring, as everything is contained in one digital container.
1 Like
And it also seems to me that such a system is inherently non-scalable. What happens when the business with 30 clients grows to 100? 200? Or more?
Why would it be inherently unscalable?
For one, having databases doesn’t require you have them constantly open. For two, you’d likely be talking about many small databases.
1 Like
I don’t mean unscalable in the sense of computing capacity - I agree with you it would be workable there and in fact maybe even a reduction in CPU needs.
What I am referring to are all the housekeeping tasks associated with setting up a database - especially if the info is shared among multiple computer seats/staff members, as is likely for a business with multiple clients.
On each computer synced to the database you would need to add the database to the sync list when it is created. If the database is encrypted, then you would have to distribute the encryption key to those with a need to know since using the same encryption key would defeat the aspect of keeping each database private.
If you have any non-global Smart Rules or Smart Groups, you would need to create those every time you create a new database.
If you use the web server, you would need to update access credentials for every new database you add.
All that said, you may be reducing rather than increasing privacy. In many legal or medical or business situations, simply knowing that someone is a client is a major secret to be retained. That can be done if clients are Groups within an encrypted database. But if you make each client a separate database, naming the database with the client name makes it obvious for someone browsing the computer to see the list of all clients. So instead you would need to assign a client number; that would make it a chore to find and open the database of a specific existing client, especially as the list grows to hundreds or thousands of clients.
And finally - what happens if you need to remove access to a particular client’s file such as if a client gets some publicity and data needs to be locked down to be accessed only by the principal of the company? If everything is kept in a small number of databases, it would be fairly simple for the Principal of the company to move that data to a “Sensitive” database that is not synced to others. But if the databases are individual per client, it would not be easy for the Principal of the company to remotely remove the case from all of the other computers synced to that data.
I think the risks you describe @rkaplan are valid, but do not outweigh the advantage of segregating client data in client-specific databases. Different matters handled for the same client can be in the same database, of course.
I believe the best practice for managing the risks would be too have a single administrator responsible for database lifecycle (initiation to archiving) in the office. If the administrator has a checklist and is the only person permitted to create databases for staff, then the firm can control security, dispersion, sync, etc. For small firms, this might be overkill.
Database creation with standard groups etc. can be handled with templates. For example, the technique used in the built-in Data > New from Template > Registers templates can be mimicked and far more complex than the sample templates. Same for smart groups, tags structure, etc. An administrator can also create a script to make and pre-populate a database, etc.
So, with simple-to-setup automation the burden you’re concerned about can be mitigated.
1 Like
Interesting thoughts - I guess the best approach depends on the nature and size of each firm.
I do have a question which evolved from this discussion. Assuming I keep each client in its own Group, I am pondering if I should Lock each group which I am not actively working on - that would be a double-check against accidentally editing or deleting data for the wrong client. I see that a Lock on a specific file prevents editing but does not prevent deleting it. I cannot figure out what effect Lock has on a group.
Would it be possible for a Lock on a group to (a) Prevent deletion or renaming of the group; (b) Prevent deleting or adding documents inside the group; and (c ) Prevent editing or renaming files inside the group
If this were to be possible, I could probably then write a smart-rule to find any group not modified within the past X days and Lock the group. That would help reduce the risk of accidental cross-contamination or accidental editing of data for the wrong client.
Locking prevents deletion on emptying the Trash or modifying a file’s contents. It doesn’t prevent renaming.
The user interface lets you “lock” a group, but I cannot see any effect that has. Is that correct?
Would it be possible for locking a group to prevent editing or deleting any contents of the group?
Locking a group prevents it from being deleted when emptying the trash.
No, it is currently not possible to prevent editing the contents of a locked group.
The deeper I dive into DT3 the more I miss the three pane view. It really suited how my work flows across multiple databases. To be frank, the work arounds being discussed are not clicking for me. I strongly hope that the dev team brings it back.
In the interim I have pulled my business data from DT and reverted to files and folders. It is just simpler right now, especially since I am stuck in MS Office for business reasons (cross platform with clients and colleagues) and MS won’t support features like autosave unless you are using onedrive.
I am not giving up on DT as it is early in the version 3 game, but DT is no longer a primary tool. On a more global sense the whole MS office, various proprietary cloud services seeking lock in, Catalina, you name it mess has me pining for building up a simple linux system and pure text files!
8 Likes
As the original original discussion moderator…a lot of folks are missing the essence of the subject matter, whatevs! I’m still stuck with tens of thousands of objects structured based on a 3 Pane view years of workflow. This is too big a feature to get screwed over, a matter of principle for me, and @RCK I have also moved on.
5 Likes
Many thanks for this suggestion @korm. I realise that it was meant for a particular use case, but for a while I thought this set up might be a way to work around most of the shortcomings inherent in the lack of a 3 pane view. But after playing around with it, a number of problems remain, e.g. this view does not allow you to reveal more than one set of subfolders at a time, or indeed to view the contents of more than one folder at a time. Thanks anyway!
A lot (all?) of the points made in your original posts, which you illustrated very effectively last May, are pertinent to most of the users on this thread who miss 3PV. Whether it’s organisation of data or navigation of databases, 3PV workflows have been broken or heavily disrupted.
The devs have said they are working on improvements to the sidebar, so there is hope. However, my reading of the forums over the past two months suggests that the devs have no intention of bringing 3PV back, so it will be interesting to see what the improvements are. Keyboard navigation of the sidebar is one; multiple folder selection is another.
@stanblues999, can I ask what you have moved on to?
1 Like
sorry, but you still don’t get the point after the many complaints. it is about constant vertical scrolling in the sidebar. you can’t isolate a group in a new pane for easier navigation. it is just not possible with dt3. you really should listen to your longterm users with this …
8 Likes
that’s exactly what i have done … now relying on icloud. i think it is more realistic to wait for spotlight in icloud for better search than for 3pane view in dt3 … sync and sharing is easier and faster with icloud …
after that, the only thing i really will be missing from dt is page specific file linking … but well, i’ll get back to plain old citation techniques … at least they will be future proof forever …