DT3 auto group/auto classify?

That doesn’t necessarily imply that they simply removed a menu option, or something equally similar. I don’t have any inside knowledge and we won’t know. unless and until they decide to discuss it further

Features are added and removed all the time so I find it highly unlikely that the difficulty of doing so is the issue here And if it was it would’ve been easy to say “well guys we would like to put it back but it’s too hard” but that’s not iat all what they said. Anyway, yes the decision is up to them and the way it’s been handled so far is pretty shocking to me as I think I have made clear. I have been using personal computers since their inception and I cannot remember a single instance such as this where a software feature has simply been summarily removed on the grounds that “not enough people want it.” Every developer that I am aware of would rather keep a feature is in as long as some customers are using it. This is “Swiss army knife” software after all.

I’d also be interested in a more technical explanation of why they couldn’t just keep this feature. I understand the extra cost of customer support, but the suggestion to rename it to something like “Group Similar Items” with some kind of warning seems to be a good idea.

One of the main reason I use Devonthink is the AI and Auto Group was one of the best example of this.

7 Likes

I am not a native English speaker, so if I misunderstood anyone, please accept my apology in advance.

AFAIMA, the developers from DT always patiently listen and collect suggestions and weight the priority of tasks. Some messages in this thread are creating a perception (at least for me) that personal wishes are being forced down to the throat of others, and with a sense of urgency that outweighs the importance of bug-reporting and bug-fixing.

Perhaps those with more specific needs would consider waiting for the official release of DT3 and let other users who do not participate in the beta testing to make a judgement. If any specific needs are becoming salient and are consistent with the design philosophy and core competence of DT, all we need is be patient and to sort out workarounds while waiting. When needs/reasons are already clearly expressed and explained, perhaps it is unnecessary to keep repeating the same message.

We don’t use a stick to reward such a fantastic and hardworking team for giving us such a unique product. We encourage further development by purchasing or upgrade the license.

Just my 5 cents.

9 Likes

How exactly is anything “being forced down your throat”…if this thread is not to your liking,why are you reading it?

And so what…we are supposed to wait for the official release after which it will be impossible to get any changes made? This makes no sense at all.

And finally, if you don’t feel the same “sense of urgency” then you don’t need this feature as much as those of us who do. Please don’t lecture us on how to express our concerns. Again, why are you on this thread if Autogroup is not an urgent issue for you?

Every feedback is welcome as long as it’s constructive and not offending (e.g. polemics, rants etc. don’t help anyone), it’s not up to you to decide who can join this discussion, especially as he only tried to be helpful (and ngan is a very nice & helpful guy).

2 Likes

“Some messages in this thread are creating a perception (at least for me) that personal wishes are being forced down to the throat of others, and with a sense of urgency that outweighs the importance of bug-reporting and bug-fixing.”

Interesting, these comments of his are are pretty aggressive and would appear to be directed towards me and suggesting that I am the one who is not welcome to make them yet you reprimand me? I was simply asking him if he felt that way and didn’t really care about the feature under discussion why he was on the thread in the first place? How is this telling him that he is not “welcome?” How exactly was he being “helpful”…what exactly did he offer to resolve this issue other than suggesting we not criticize you?

In fact, I am now the one who feels unwelcome with the inference that I am engaging in “polemics and rants” when I am simply a long-standing customer who is actually making concrete suggestions to get this issue resolved. Perhaps if instead of this kind of response you would direct your energies to helping us solve this issue it would go along way to resolving the feelings here. A good start might be a response to the suggestions I have been making about renaming the feature. Also nice would be a single word of understanding or support for the position some of us are in at the moment.

I honestly think you’re being over-sensitive here. I don’t see anything in @cgrunenberg responses that’s aimed at you (or anyone else for that matter) - he’s simply reminding us what kinds of content are not acceptable here.

It’s clear that you feel very strongly about the issue of auto-classification, and you’ve seen some comments from others who agree with you about its value (including me). But we have to recognise that we’re not the only people here and that there are others whose issues are more important to them than this is to us. And we also have to recognise that this is work in progress and allow DT folks time to fix the issues in the beta so they can release it and turn their minds to new/returned features.

6 Likes

"I honestly think you’re being over-sensitive here. I don’t see anything in @cgrunenberg responses that’s aimed at you (or anyone else for that matter) "

Always possible except that he singled me out by name.

“we have to recognise that we’re not the only people here and that there are others whose issues are more important to them than this is to us”

OF course but this is a thread about the Auto Group feature.I have never suggested that this should be the only concern or that even a single second of further development time be devoted to the feature. All I am asking is that it not be removed and have suggested a rename that could avoid the issues with the feature as they have been explained. If @cgrunenberg had taken the same amount of time he took to single me out in that comment to at least respond to the suggestion I would not be so irritated.

“we also have to recognise that this is work in progress and allow DT folks time to fix the issues in the beta so they can release it and turn their minds to new/returned features.”

Of course but yet again, there is no indication what so ever that Auto Group is returning. The only thing we have heard to date is “not many people use it” and “it doesn’t work as expected.”

Maybe I seem overwrought about all this but please understand this is not a hobby, an avocation, or any other form of optional activity. This is my livelihood and I use Auto Group in my business almost every week of my life and have been doing so for almost 15 years. I cannot replace it and its loss would be a blow to my ability to what what I have been doing. All I have been asking for is a single expression of acknowledgement and a willingness to consider suggestions that address what the devs have said the issue really is. Instead, Iwhat I got today is a statement about what a “nice and helpful guy” my critic is. Its just a bit too much for me and not my normal experience with devs and loyal customers.

Lol…yours is a kind, sympathetic, and understanding response. I would be more likely to kiss you than hit you, metaphorically that is :slight_smile:

But the most painful thing is that it is so unecessary when a simple rename would fix everything and would likely be a valid featured by old and new customers when they understand what it really does. Why on earth would you remove a functioning bit of AI which is billed as central to the software, particularly when there is nothing wrong with it other than customers who don’t understands how it is supposed to work?

Oh well, I’ve already been accused of obsession so I better stop here.

Thanks again for understanding.

1 Like

I am reading this forum regularly because I admire and expect those experienced members who always voluntarily offer solutions for a new/existing problem. But the prerequisite is perhaps a clear description of issues and specification of an acceptable solution. There are many members with research background, with curiosity and engineering mind who might and could suggest workaround by using existing functions or by scripts or by integrating with the external apps to proxy the way you might need. I think that if frustration can be turned into a series of equations for solving a specific problem and to prioritising which equation/s are most salient, it is more likely for others to offer help. The more details you give in describing your flow/method of info management (e.g. source/nature of info/filetype/pages/words/language of incoming files), how you want to group it (e.g., by keywords/hits of keywords/connected keywords/sources/dates/some format of data/some metadata), and what outcomes are acceptable, the more likely there will be workarounds. Obviously, it depends on the time of other members. But there are 12 months or more of maintenance period in DT2.n, and IMHO that’s a very reasonable grace period for any app.

Another alternative is to PM the developer to understand the general concept of auto-grouping, and if you know how it works, then there is likely a way to get an approx/acceptable outcome by scripts.

I see DT2/3 as an integrated platform that consolidates many handy Unix functions and proprietary algorithms into the form of scriptable commands, many useful scripts, and stock functions that can manipulate and processes the items within the database. If you look into the script library of DT3, all core data manipulation command/elements such as to get concordance, compare, classify, perform smart rule, etc. are available at your disposal.

If the solution is staying with DT2.n and that also seems to be perfectly acceptable. DT2.n is already a best-in-flexibility, user-friendly, organic, and stable info-management app I have ever encountered in my 30 years of search for a perfect info-mgt app. Finally, (I am guessing again) given all the new functionalities in DT3, it seems that there are huge chunks of new codings in DT3 and is definitely not merely a face-lifting job. So, retaining any/certain functions very likely involving a lot of re-coding in different areas and can only be justified by notable demands.

Again, just my 5 cents.

Sorry but I There is no way I can envision doing this… There is no “specific problem”. What I need is exactly what I had before – a quick and dirty way to group similar documents into folders so I can proceed further without wasting time.

“Another alternative is to PM the developer to understand the general concept of auto-grouping, and if you know how it works, then there is likely a way to get an approx/acceptable outcome by scripts”

Did this already and the developer said it would be “very difficult” And showed no interest in doing it.

“If the solution is staying with DT2.n and that also seems to be perfectly acceptable.”

It’s not acceptable… Like everybody else I would like to benefit from the development of the app and not be constrained forever to an old version

“retaining any/certain functions very likely involving a lot of re-coding in different areas and can only be justified by notable demands.”

I Highly doubt it and that’s not what the developers are saying.

Look, I appreciate what you were trying to do but I’m already pretty much burned out on discussing this. In my opinion the response to these concerns has been appallingly bad- unoncerned, and unresponsive. Sorry but that’s the way it seems from this end. I’ve expressed how catastrophic this is for me made a suggestion on how to easily solve the problem and the only response I’ve got is to tell me that somehow I’ve made you feel “unwelcome “here. I’ve been involved with computing for over 40 years and I don’t remember very many experiences like this involving a loyal and long time-supporter of a product. At the minimum all I ever asked for is a word of support from the developers showing that they least understand the concern and I’ve been met with glaring silence.

So be it… I need Auto Group to come back and that’s the end of it for me. I don’t have any sense that I’m being listened to by the developer so I see no sense in carrying on this conversation further But thank you for trying to help, that’s more than They have done

1 Like

Thanks for the explanation and I’m sure that you will be able to find an alternative here or elsewhere. No trying to sound stupid, but there is always a solution waiting to be found and you may also change ur perspective in the journey.

Not at all stupid but wise advice, in general at least. Problem is that I cannot afford a major disruption in workflow as clients have no understanding for life journeys :slight_smile:

I’m a bit late to this exchange, but I join in the request for the return of Auto-Classify in DT3. Calling it “Group Similar Items” also sounds like a good idea.

I was looking forward to using Auto Group/Classify to start my next project. My work is problem-centered: it follows a question across disciplinary boundaries. Auto-classify was a big help, and even its occasional oddball connections could lead to unanticipated insights. It is, as Eric pointed out, was one of the AI features “that gave DTPO its reputation.”

“Auto Classify are some of the intelligent functions that gave DEVONthink Pro Office its reputation. It tries to find similarities between the selected documents and then group them accordingly. This is most useful, for example, when trying to create order and structure out of a few hundred documents randomly clipped from web sites.”

Eric Böhnisch-Volkmann. DEVONthink Pro Office Documentation (Kindle Locations 1514-1516). Kindle Edition.

3 Likes

Exactly!

Here is my suggestion for the documentation on “Group Similar”:

This feature is designed to be a “pre-processing” step for large documents sets and performs a textual analysis on the selected documents. attempting to place them in folders according to similarity. Depending upon the nature of the documents selected, the resulting groups may be composed of very similar documents, duplicates, documents from the same source, same language documents, and more. It is possible that only a portion of the document set will be placed into folders with the remaining documents listed individually. If the resulting folders are not found to be helpful, the process can be easily reversed by re-selecting the folders and clicking on “Ungroup.”

Thoughts?

OK, I guess there is no more interest in this topic or everything that can be said has been said. I can only hope that the developers see reason and re-activate auto group.

2 Likes