Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Clément]

It is easy to add them, the question is where. A new tab for each ? Or add them both in an existing tab (e.g. Social) ?

To make sure I don't forget anything. Performer skills are:

• Dance
• Music
• Singing
• Keyboard instrument
• Stringed instrument
• Wind instrument
• Percussion instrument

Scholar skills are:

• Critical thinking
• Logic
• Mathematics
• Astronomy
• Chemistry
• Geography
• Optics engineering
• Fluid engineering

Anything else to add?

[Hartsteen]

Great, thank you! The social tab seems appropriate to me.

Anything else to add?

"Poet", "Writer" and "Wordsmith", I guess.

[Clément]

I'll use the list from here. That add "Speaker" too.

[Clément]

I also split the social skills using the Wiki categories. I posted a screenshot in the pull request comments. Does that look good?

[PatrikLundell]

I think it looks very nice. I also think it makes sense to have the "old" record keeper etc. skills on the same tab as the "new" ones.

[Hartsteen]

Superb!

[feelotraveller]

I also split the social skills using the Wiki categories. I posted a screenshot in the pull request comments. Does that look good?

Very much in a hurry so I will look further later.  Been a while since I did this though.

Performance category looks very good.  I had "StoryTeller" instead of "Speaker" in my version.  Just flavour...

The scholars are a bit different.  I had Philosopher which I think = Critical Thinker but also Historian and Naturalist. I also folded the two engineers together but that's just me .  If I remember correctly Writer (for scholars) and Wordsmith (for Poets) are not roles but contribute to other roles.  I also made a Scholars entry (Studying, Teaching, Writing) for general assignment, and a Scribes one that was mainly redundant.

Really got to run, white rabbit moment, let me know if you have questions, it will probably be a day or two until I'm back...

[Clément]

They are skills not roles. The names are from the game, there is not much to debate (but I could have made a mistake while copying them).

I did not add any role related to this (and I don't think there is currently any). You can make detailed suggestions when you have the time. I did not play enough with this to know which roles would be useful.

Edit: Since you are speaking of naturalists, I see I am missing the tracker skill. I dismissed it as an adventure skill, but it is also used for scholars.
Edit2: Is "reader" of any use in Fortress mode?

[Fleeting Frames]

Naturalist use Tracker, Observer and Record keeping. The skills and topics aren't 1-1 match or even necessarily same research path from embarking with same skills.

Reader is useful for finding out who has read how many books and I think that's it?

[feelotraveller]

   Yes skill not roles.*

I'm glad the Naturalist comment had some redeeming merit.

There was some speculation that Reader helped with the transmission of values via books but I'm not sure if this was ever more than idle speculation.  Otherwise the consensus seems to be that there is no known use for reader in fortress mode.

I was tempted a couple of months ago to make another pass through the roles, and probably will be in a couple of months more when I get some time.  I already have a set of roles for the performance and scholar positions and a bunch of other changes/suggestions, probably the most controversial would be to split the roles into two pages.  In the meantime if you are interested... there's a couple of things I wonder if could be looked at (I don't have the skills...)

a) preferences for cloth and paper are currently not distinguishable in the roles, e.g. a dwarf with a preference for kenaf paper will be considered a better clothesmaker
b) the role suitability calculations currently use every role.  can we get the option/ability to specify which roles to use?

In the meantime I built a linux-64 version of DT from your git repository. At least for the basics of connecting and reading dwarfs from an old save there were no problems.  When I find time to play again I'll give more feedback.  But I wonder about the LaTeX?  I did not build the manual since 400mb roughly for the software seems excessive to me (and I already have a copy).  Would it not be easier to just include a PDF of the manual - or what am I missing?  (You may already know - the author is still around and has dropped by a couple of times checking on how it's going, if there is some thought of an updated version...)

Anyway, all the best, it is great to see someone with enthusiasm for maintaining/developing Dwarf Therapist; don't let me put you off. 


*I would add a note to self not to post when pressed for time and exhausted, but pressed for time and exhausted self would not bother with silly notes...

[Clément]

a) preferences for cloth and paper are currently not distinguishable in the roles, e.g. a dwarf with a preference for kenaf paper will be considered a better clothesmaker

When not doing exact matches, DT only compares material flags. A dwarf that likes linen paper has material PLANT:FLAX:THREAD (flags ITEMS_SOFT, THREAD_PLANT, STOCKPILE_GLOB_PASTE), the same as a dwarf that likes linen fabric. I think the material state need to be included in the comparison (PRESSED for the paper, SOLID for the fabric).

b) the role suitability calculations currently use every role.  can we get the option/ability to specify which roles to use?

You want the ability to specify the list of roles used in a dwarf details/tool-tip, did I understand correctly?

In the meantime I built a linux-64 version of DT from your git repository. At least for the basics of connecting and reading dwarfs from an old save there were no problems.  When I find time to play again I'll give more feedback.  But I wonder about the LaTeX?  I did not build the manual since 400mb roughly for the software seems excessive to me (and I already have a copy).  Would it not be easier to just include a PDF of the manual - or what am I missing?  (You may already know - the author is still around and has dropped by a couple of times checking on how it's going, if there is some thought of an updated version...)

I re-added the manual compilation after the last release, I will include it in the builds for the next release (but I don't provide any linux build, so that won't make a difference for you). The manual is available as a stand-alone download on DFFD (clicking the "manual" button in DT opens this url). I don't know of any changes (according to git there was no change for 3 years), the version number matches at least.

[feelotraveller]

Sounds like you've got it covered for the paper/fabric distinction.    (Fallback workaround could be to check the text strings of preferences for 'paper' or 'cloth' matches, but that is so ugly...).

For b) no, I have not been clear.  The roles calculation spits out a number between 0 and 100 for how well each dwarf is suited for each role.  Problems arise when adding custom roles.  Say I add a role Carpenter+, which is a revision of the attributes, preferences, skill(s) I want in finding a carpenterdwarf.  Now the data set for suitability includes both Carpenter and Carpenter+ and that has knock on effects as to the reported suitability for other roles - basically if the roles are minimally different then the weight of that role as a data point has just been doubled.  (The calculation has always been opaque to me but includes various processes of 'normalisation' 'shaping' etc.)  I would like to be able to specify which roles are to be used for the roles calculation.  So following the example above I could tell DT not to use Carpenter as part of the data set for roles calculation but only Carpenter+.  Or to put it another way (since I often struggle with clarity of expression):  can we get the feature to ask DT to consider the suitability of dwarfs with respect to a limited - user specified - set of roles?  Taking a concrete example (not that I would use it this way...) - lets say we have just embarked and know we want 7 jobs done and don't care whatsoever about future plans.  We want a Miner, a Mason, a Woodcutter, a Fisher, a Herbalist, a Carpenter, and an Animal Trainer.  So we want tell DT to limit the roles calculation to just those 7 roles (and not be worried by how suitable the various dwarfs are as Strand Extractors or Potters).

It would be good to include a note on the github page about the availability of the manual from within DT (or even a link to the download on dffd) so that new users compiling it for themselves don't feel obligated to install latex just to get the manual.

[Clément]

Sounds like you've got it covered for the paper/fabric distinction.    (Fallback workaround could be to check the text strings of preferences for 'paper' or 'cloth' matches, but that is so ugly...).

No, I don't think partially matching strings would be simpler.

For b) no, I have not been clear.  The roles calculation spits out a number between 0 and 100 for how well each dwarf is suited for each role.  Problems arise when adding custom roles.  Say I add a role Carpenter+, which is a revision of the attributes, preferences, skill(s) I want in finding a carpenterdwarf.  Now the data set for suitability includes both Carpenter and Carpenter+ and that has knock on effects as to the reported suitability for other roles - basically if the roles are minimally different then the weight of that role as a data point has just been doubled.  (The calculation has always been opaque to me but includes various processes of 'normalisation' 'shaping' etc.)  I would like to be able to specify which roles are to be used for the roles calculation.  So following the example above I could tell DT not to use Carpenter as part of the data set for roles calculation but only Carpenter+.  Or to put it another way (since I often struggle with clarity of expression):  can we get the feature to ask DT to consider the suitability of dwarfs with respect to a limited - user specified - set of roles?  Taking a concrete example (not that I would use it this way...) - lets say we have just embarked and know we want 7 jobs done and don't care whatsoever about future plans.  We want a Miner, a Mason, a Woodcutter, a Fisher, a Herbalist, a Carpenter, and an Animal Trainer.  So we want tell DT to limit the roles calculation to just those 7 roles (and not be worried by how suitable the various dwarfs are as Strand Extractors or Potters).

I am even more confused now. I am not a advanced role user, so I may be missing something. The places I look for role suitability are dwarf details or the role tab (grid view). For dwarf details, a global filter list could be added in the options. For the role tab, you can replace it with a custom grid view with only the columns you are interested in. Or do you want a new dialog or dock?

Anyway, I need to look more at how the roles work internally. And the custom role dialog looks like it needs some improvements.

It would be good to include a note on the github page about the availability of the manual from within DT (or even a link to the download on dffd) so that new users compiling it for themselves don't feel obligated to install latex just to get the manual.

I added the DFFD link in the Readme. As for the "Manual" button, it is clearly visible in the tool bar (and the help menu), and the best place to explain it would be... the manual (sections 5.5 and 5.7 are missing it).

[feelotraveller]

Sounds like you've got it covered for the paper/fabric distinction.    (Fallback workaround could be to check the text strings of preferences for 'paper' or 'cloth' matches, but that is so ugly...).

No, I don't think partially matching strings would be simpler.

Yes I agree totally.

For b) no, I have not been clear.  The roles calculation spits out a number between 0 and 100 for how well each dwarf is suited for each role.  Problems arise when adding custom roles.  Say I add a role Carpenter+, which is a revision of the attributes, preferences, skill(s) I want in finding a carpenterdwarf.  Now the data set for suitability includes both Carpenter and Carpenter+ and that has knock on effects as to the reported suitability for other roles - basically if the roles are minimally different then the weight of that role as a data point has just been doubled.  (The calculation has always been opaque to me but includes various processes of 'normalisation' 'shaping' etc.)  I would like to be able to specify which roles are to be used for the roles calculation.  So following the example above I could tell DT not to use Carpenter as part of the data set for roles calculation but only Carpenter+.  Or to put it another way (since I often struggle with clarity of expression):  can we get the feature to ask DT to consider the suitability of dwarfs with respect to a limited - user specified - set of roles?  Taking a concrete example (not that I would use it this way...) - lets say we have just embarked and know we want 7 jobs done and don't care whatsoever about future plans.  We want a Miner, a Mason, a Woodcutter, a Fisher, a Herbalist, a Carpenter, and an Animal Trainer.  So we want tell DT to limit the roles calculation to just those 7 roles (and not be worried by how suitable the various dwarfs are as Strand Extractors or Potters).

I am even more confused now. I am not a advanced role user, so I may be missing something. The places I look for role suitability are dwarf details or the role tab (grid view). For dwarf details, a global filter list could be added in the options. For the role tab, you can replace it with a custom grid view with only the columns you are interested in. Or do you want a new dialog or dock?

We are thinking of it in different ways.  In your language I want a global filter list (but I want it to have greater functionality than what I understand you to be suggesting). Let me expand a bit...

A simple filter could remove the extra roles from the dwarf details, so yes for this.  And yes merely by having a customised grid view I can restrict the display to only certain roles.  However ALL the roles are still being used to calculate the numbers.  Without knowing precisely what the code for determining these numbers actually does - though I am sure it is complex - I can tell from using DT that the data matrix (X dwarfs by Y roles) responds to the addition and refactoring (i.e. changing variables and/or values) of roles in complex ways - in short, adding a role impacts on the suitability of dwarfs for other roles (=the reported number).  Basically each dwarf is being compared for suitability to each role even the ones that we are not using (even if we have a global filter and a restricted customised gridview).  These dwarf/role (x/y) points are then 'shaped' (normalised, or otherwise statistically manipulated) in both dimensions, that is over the set of dwarfs and over the set of roles.  I really don't think we want to go the potentially dangerous route of allowing users to delete default roles but that would be a solution of sorts.  I would like to be able to run the calculation of suitability over a limited set of roles rather than having to use all of them. So the extended functionality I would like for the global filter list is for it to exclude the roles both from display (dwarf details) and from use in the calculations (reported number production).

I hope this is becoming clearer.  I can post tomorrow but then will not be able to for some time.

Anyway, I need to look more at how the roles work internally. And the custom role dialog looks like it needs some improvements.

Looking forward to custom role dialogue improvements.    In a couple of weeks time I could make some suggestion if you are interested. 

It would be good to include a note on the github page about the availability of the manual from within DT (or even a link to the download on dffd) so that new users compiling it for themselves don't feel obligated to install latex just to get the manual.

I added the DFFD link in the Readme. As for the "Manual" button, it is clearly visible in the tool bar (and the help menu), and the best place to explain it would be... the manual (sections 5.5 and 5.7 are missing it).

Perfect. 

[Clément]

I started a fix for the cloth/paper bug: https://github.com/Dwarf-Therapist/Dwarf-Therapist/pull/9

The custom role dialog material preferences list need fixing too: adding a cloth preference from there will also match paper (and you cannot select paper preferences either).
The problem is that it generates the list from the game material list but I don't how to get the valid material states from there. Maybe from "material_common.state_name"? Getting interesting material preferences using it does not look easy. If I look with DFHack at a few of one dwarf preferred materials, I get:

Code: [Select]

<material 0:82 INORGANIC:JASPER OPAL>
ITEMS_HARD
IS_GEM
ITEMS_QUERN
Solid jasper opal
Liquid magma
Gas boiling magma
Powder jasper opal
Paste jasper opal
Pressed jasper opal

<material 38:343 CREATURE:GIANT_CHEETAH:LEATHER>
ITEMS_LEATHER
IMPLIES_ANIMAL_KILL
LEATHER
Solid leather
Liquid n/a
Gas n/a
Powder leather
Paste leather
Pressed leather

<material 37:476 CREATURE:SPIDER_PHANTOM:SILK>
ITEMS_SOFT
SILK
Solid silk
Liquid n/a
Gas n/a
Powder silk
Paste silk
Pressed silk

<material 425:29 PLANT:HEMP:THREAD>
ITEMS_SOFT
THREAD_PLANT
STOCKPILE_GLOB_PASTE
Solid hemp
Liquid none
Gas none
Powder fiber
Paste hemp slurry
Pressed hemp paper

I see both "none" and "n/a" for invalid states, and some material states use generic names (e.g. silk, leather, fiber, magma). I don't know how I should process that.

I hope this is becoming clearer.  I can post tomorrow but then will not be able to for some time.

Yes, I understand now. I did not know how the value were calculated.

Looking forward to custom role dialogue improvements.    In a couple of weeks time I could make some suggestion if you are interested.

Not really a redesign, more internal stuff you might not see: memory leaks and the way it loads data from the game. Still, suggestions are welcome.