Lexicon management?

[Jonlang]

I'm about to re-build my conlangs lexicons. The three languages are all related and share etymologies. So what I need is a means of having all three terms be searchable, with the English translations, their cognates and original proto-form. I have an old app called iDatabase but it's pretty poor. I've not really used spreadsheets or databases since around 2003 so I'm not that up to speed with them, other than simple formulae in Excel. Does anyone have any suggestions?

[Man in Space]

Let’s start with your preferred formatting for this information. What does a useful dummy entry look like for you?

[Jonlang]

Let’s start with your preferred formatting for this information. What does a useful dummy entry look like for you?

Something like 7 columns: Proto | Lang 1 | Lang 2 | Lang 3 | Root | English | Notes – which can sort alphabetically by column. I don't really need a search function, I can just scroll to find what I need in whichever language. I'm getting to a point where I can pretty much know what a cognate should be now but sometimes there are surprises.

I don't have Excel but I do have macOS Numbers which should be able to do this, right?

Unless anyone has a better tool they have for this sort of thing, or know of any useful templates out there?

[Man in Space]

I just use GDocs.

[bradrn]

I like using SIL Toolbox for my languages. It’s pretty flexible, and can do a lot. Plus the underlying format is plain-text, so if you need anything extra you’re not restricted to what the software gives you.

[Emily]

yeah i also just use a spreadsheet. (currently stuck using libreoffice until i get around to pirating excel on the new computer.) generally i keep a separate spreadsheet for different languages, but my current modern gothic project has multiple stages of the language so i'm trying to keep track of one etymon per row. i have different tabs for different parts of speech (below is a screenshot from my nouns tab) but that might be overly nitpicky on my part. i have conditional formatting graying the gloss if it matches the gloss from the previous stage (ph1 = phase 1 = "middle gothic", ph2 = "early modern gothic", ph3 = "modern gothic"), and highlighting certain cells in pink if there are duplicates (e.g. ph1 /gwás/ is the lemma for both "cow" and "speech, address")

[WeepingElf]

Relevant.

Indeed, I follow this advice and keep my conlangs in simple hand-coded HTML files - thereby also not needing any conversion when I am going to put them on my web site.

[bradrn]

Relevant.

Indeed, I follow this advice and keep my conlangs in simple hand-coded HTML files - thereby also not needing any conversion when I am going to put them on my web site.

This is why I use SIL Toolbox — its native format is MDF, which is plain text (somewhat reminiscent of LaTeX, but a lot simpler). Indeed, it’s trivial to edit the dictionary without using Toolbox. The software just provides a convenient way to manage that plain-text dictionary.

(For a sample, I recently uploaded the MDF-format dictionary for my reconstruction relay language: http://bradrn.com/files/akana-2023-rela ... ionary.mdf)

[chris_notts]

Relevant.

Indeed, I follow this advice and keep my conlangs in simple hand-coded HTML files - thereby also not needing any conversion when I am going to put them on my web site.

This is why I use SIL Toolbox — its native format is MDF, which is plain text (somewhat reminiscent of LaTeX, but a lot simpler). Indeed, it’s trivial to edit the dictionary without using Toolbox. The software just provides a convenient way to manage that plain-text dictionary.

(For a sample, I recently uploaded the MDF-format dictionary for my reconstruction relay language: http://bradrn.com/files/akana-2023-rela ... ionary.mdf)

I tried it in the past, but I find the interface a bit clunky and painful. My own approach now is often a mix of spreadsheets and Python code that does some analysis (phoneme frequencies by position, closest word, ...).

[bradrn]

Relevant.

Indeed, I follow this advice and keep my conlangs in simple hand-coded HTML files - thereby also not needing any conversion when I am going to put them on my web site.

This is why I use SIL Toolbox — its native format is MDF, which is plain text (somewhat reminiscent of LaTeX, but a lot simpler). Indeed, it’s trivial to edit the dictionary without using Toolbox. The software just provides a convenient way to manage that plain-text dictionary.

(For a sample, I recently uploaded the MDF-format dictionary for my reconstruction relay language: http://bradrn.com/files/akana-2023-rela ... ionary.mdf)

I tried it in the past, but I find the interface a bit clunky and painful.

You’re not wrong! But I find that for my purposes, it works well enough. And being plain text makes up for a lot, since it’s so easy to process via other means.

[chris_notts]

You’re not wrong! But I find that for my purposes, it works well enough. And being plain text makes up for a lot, since it’s so easy to process via other means.

So are spreadsheets though! You can manually export to CSV, but I just use pandas.read_excel in my Python code to read in the summary table so I can calculate the stats and transform into a LaTeX dictionary.

I guess it's the programmer urge to just hack your own thing together whenever the out of the box solutions don't quite do what you want.

[bradrn]

You’re not wrong! But I find that for my purposes, it works well enough. And being plain text makes up for a lot, since it’s so easy to process via other means.

So are spreadsheets though! You can manually export to CSV, but I just use pandas.read_excel in my Python code to read in the summary table so I can calculate the stats and transform into a LaTeX dictionary.

I guess it's the programmer urge to just hack your own thing together whenever the out of the box solutions don't quite do what you want.

What I dislike about spreadsheets is that everything is stuck in the grid of single cells. So if I, say, want to write a definition longer than a single line, that becomes difficult. It gets even more difficult if I want to give words more than one sense, add explanatory comments, specify paradigms, or in general do anything which doesn’t fit into a grid neatly. Plain-text formats like MDF solve all these problems quite simply.

[tlacamazatl]

yeah i also just use a spreadsheet. (currently stuck using libreoffice until i get around to pirating excel on the new computer.) generally i keep a separate spreadsheet for different languages, but my current modern gothic project has multiple stages of the language so i'm trying to keep track of one etymon per row. i have different tabs for different parts of speech (below is a screenshot from my nouns tab) but that might be overly nitpicky on my part. i have conditional formatting graying the gloss if it matches the gloss from the previous stage (ph1 = phase 1 = "middle gothic", ph2 = "early modern gothic", ph3 = "modern gothic"), and highlighting certain cells in pink if there are duplicates (e.g. ph1 /gwás/ is the lemma for both "cow" and "speech, address")

I use a similar technique. It's most helpful for sorting, finding, and editing. I then copy over (with copious editing) to my word doc (when there are changes) which I republish every month or so as the latest version of my grammar.

[KamareKitsune]

Probably the main reason I'm still using spreadsheets is that I highly prioritise being able to filter, not just search. And speaking of search, being able to search in just the translations is rather nice, especially if you have a prose discussion as well.

[hwhatting]

Probably the main reason I'm still using spreadsheets is that I highly prioritise being able to filter, not just search. And speaking of search, being able to search in just the translations is rather nice, especially if you have a prose discussion as well.

Same for me. I use Excel a lot in my day job anyway, so it's a software I feel at home with.

[xxx]

in 3SDeductiveLanguage(1Sense=1Sign=1Sound),
I only have a hundred or so semantic primes...
it all fits on a credit-card sized document,
that contains the second equality (1Sign=1Sound) in my wallet...

lexicons are forbidden to avoid approximations,
whereas each agglomeration of primes is its own definition...
as well as texts storage to avoid repeating previous compositions,
rather than (re)finding them when necessary...

everything is memorized,
including the morphological system
logic prevails,
so that on-the-fly composition is possible...
so except my memory card for remembering the sound of the least-used idiograms,
everything is just stored in my brain...

[WeepingElf]

Probably the main reason I'm still using spreadsheets is that I highly prioritise being able to filter, not just search. And speaking of search, being able to search in just the translations is rather nice, especially if you have a prose discussion as well.

Nice semantic extension of the English word prose: 'text not in verse' > 'text outside lists or tables'.

[KamareKitsune]

Probably the main reason I'm still using spreadsheets is that I highly prioritise being able to filter, not just search. And speaking of search, being able to search in just the translations is rather nice, especially if you have a prose discussion as well.

Nice semantic extension of the English word prose: 'text not in verse' > 'text outside lists or tables'.

Honestly, that's not something I've ever considered before! I've just thought of it as "normal running text", and so lists and tables don't count, since they're not running text.