LF Workflow - what can it look like with Flex and PT9?

Language Forge
1

Hello Chris and Alex,

The new improved LF looks really good. I am trying to picture how (and if :slight_smile: ) to implement LF in our current workflow and how it will fit in. So this is a translation team also skilled in lexicography working with Paratext using a link with flex. So PT9 and Flex are set up with send receive and some members of the community are willing to engage with the dictionary on their laptop or mobile device. How would you propose our workflow to look like?

Further questions:

  • Is it possible to offer certain users a limited view and commenting options?
  • Can community users be shown a subset of the complete dictionary, but the translation team the full set ?
  • the LF documentation says " … LF handles a variety of lexical fields. …" Is there a list of all the fields it can handle, and of those it cannot ?
  • Is it possible to “Freeze” a certain subset temporarily for major maintenance ? (i.e. bulk edits in flex)
  • can commenting be done off-line (temporarily) ?
  • can users with “view only” be presented with a DAB look-a-like or flex dictionary preview kind of look ? Asking because is it wise to just have a limited amount of knowledgeable people use LF and for view purposes distribute the DAB app version …?

Some points for improvements at first glance (in the help windows)

  • the “Start Language Forge Tour” does not work
  • Tutorials and how-to is the same as community support
  • The what is new in the app is quite lot of information, and understood by few ?

Thanks a lot for your help.

2

Dear Bart,

Thank you for reaching out about Language Forge workflow. You’ve got some great questions - I will try and answer them one at a time below:

So PT9 and Flex are set up with send receive and some members of the community are willing to engage with the dictionary on their laptop or mobile device. How would you propose our workflow to look like?

LF is designed to send/receive with FLEx lexicon. Language Depot is where the FLEx data is stored for S/R. LF uses the same S/R data in Language Depot. To get started with Send/Receive, you will first need to create a new LF project using your Language Depot username/password. You must be a Language Depot project manager in order to do this first step.

After FLEx and LF are connected like this, I could imagine you making changes to the FLEx lexicon directly in FLEx or via Paratext (I assume there are ways to modify the FLEx lexicon from Paratext but I’m a bit light on details). At times convenient for you, you can do a Send/Receive in FLEx and LF in order to sync up changes to/from these programs.

Think of LF like a group of more casual lexicon users that all have some access to the shared lexical database. This in contrast to FLEx and PT which are installed on individual machines and each have their own offline copy of the data.

Is it possible to offer certain users a limited view and commenting options?

Yes, we have an observer role and a commenter role. For these roles, you can customize which fields and writing systems they see. This is how we originally designed LF.

Can community users be shown a subset of the complete dictionary, but the translation team the full set ?

This is an often requested feature, but we haven’t implemented it yet. Most people I meet want to be able to do this with their dictionary. Thank you for bringing it up. There are multiple way to implement this sort of thing. One of the way is to utilize the “publish in” field that FLEx already has when grouping a set of entries for display or access.

the LF documentation says " … LF handles a variety of lexical fields. …" Is there a list of all the fields it can handle, and of those it cannot ?

We are currently maintaining a document called Known Issues and Limitations which includes a list of FLEx fields we don’t yet support. Lack of support for FLEx fields is a result of limited developer resources to continue building compatibility with FLEx - not lack of desire on my part :slight_smile:

Is it possible to “Freeze” a certain subset temporarily for major maintenance ? (i.e. bulk edits in flex)

This isn’t currently possible in LF. This is an interesting feature idea - I have created issue #1380 in our tracker for this and it would be great if you could help us develop the feature more there.

can commenting be done off-line (temporarily) ?

Unfortunately, no off-line features are implemented yet.

can users with “view only” be presented with a DAB look-a-like or flex dictionary preview kind of look ? Asking because is it wise to just have a limited amount of knowledgeable people use LF and for view purposes distribute the DAB app version …?

This is also something a number of people have asked for. We are making plans with our re-write of the application to produce a view-only view similar to the look of DAB and Webonary. Thank you for bringing it up as I understand it is a priority.

Some points for improvements at first glance (in the help windows)

Thank you for pointing out the discrepancies in the Help Menu links. I’ve changed the links in the menu around to hopefully be more helpful, and renamed it to “About”, since there’s more there than just Help. Here’s what it looks like now:

image

3

Thank you for the explanations Chris. This will help us set things up well from the beginning.

One of the big challenges for us will be to train editors / commentators in lexicography first. Something we have to carefully plan. Otherwise I am afraid that the direct contact with LF will not give us the results we like to see.

About the question of offering certain viewers / editors a limited set I can see several possibilities:

  1. “filter” the sets by creating different databases in Flex and thus each LF project mirrors a subset. This will probably mean double user management, will it not ?

  2. A much better option is to have a real admin filter feature in LF allowing yes / no visibility based on certain field values. For example, on the side, we are developing specialised vocabulary for teaching geometry and sciences in schools. Each entry of specialised vocabulary has a tag in a certain field which we already use for our own ease of working with the database. We also decided to include proper names in the dictionary, allowing people to look up their own name and how to write it. These entries are also separately tagged.

Another question that came to mind is the possibility to allow for customization. As an example : We use a part of speech that is fine-grained, that distinguishes between transitive verb, intransitive, etc. the kind of fine-grained information linguists like to see. In the dictionary that we want to publish however, we will be using a much simpler part of speech, in this case just verb. In the flex googlegroups exchanges I have been advised to create a custom field for the simplified version of the part of speech.

Now I just looked up the text about custom fields in LF:
————
At this time, Language Forge does not support creating, editing, or removing custom fields natively. These changes must be made in FLEx, and in synchronization with Language Forge.

Follow these instructions:

1. All FLEx users and a Language Forge admin must do a Send/Receive. No one except the computer in step 2 may make any changes until step 4.
2. Once step 1 is fully complete, the computer adding the custom field must do Send/Receive one more time.
3. Create the custom field in FieldWorks, and do a Send/Receive after editing it. (Do NOT change or create any project data yet; it’s best if the changes in this Send/Receive ONLY include the new custom field(s) and nothing else).
4. Everyone on the team does another Send/Receive (including an administrator on Language Forge) so that they receive the field change.
5. Resume working normally.
The above protocol also applies to editing custom field properties, or removing custom fields.
—————

Will this mean that the simplified part of speech I created will appear in LF ?
This would be great … at the same time I read:
Some types of FLEx custom lists will cause a project to go on hold in Language Forge. (See Issue 976 for more information.)

Another option for us would be to swap our part of speech fields. The simplified form in the standard field for part of speech and the fine-grained form in the new custom field. What would you propose ?

4

Hi Bart,

You might find the introductory lexicography course () from Dictionary & Lexicography Services helpful (even though it’s in English). David

5

Hi David, yes I saw the site. The section “importing Toolbox Databases into FLEx” looks promising once it is finished. Do you (pl) have a time frame on that ?
Bart.

6

I sent this message to Bart off-list. Now I’m posting it here in case it is helpful to anyone else.

At this point, the content of the “Importing Toolbox Databases into FLEx” course is at a stage where it provides sufficient information to make it dangerous. By that I mean that I needed a lot of personal help from an expert to apply it successfully to the import project I was working on. If I had tried to import the database on my own, working only with the instructions provided in the course as it is currently, I would have failed.

Here is a place where you can find various video playlists compiled by SIL’s Dictionary & Lexicography Services (DLS): Dictionary & Lexicography Services - Videos. Included there is a series of videos on Importing SFM to FLEx. They may be of help to you if you decide you want to continue to do the import of your data yourself.

However, I wonder if you are aware that SIL’s Dictionary & Lexicography Services (DLS) provides the service of importing a Toolbox database into FLEx? (Application form is here: (Webonary – Dictionaries and Grammars of the World).) Verna Stutzman, DLS Coordinator, has assembled a team of people who have been (or are being) trained for this task, thus sparing database owners the time and effort required to master the skill. From your posts on the FLEx list, I conclude that you have become quite proficient in the use of (and manipulation of) Toolbox. I would guess then that your database might be complex enough that importing it into FLEx will not be a trivial task and might be best handled by the DLS team of experts.

Here are some of the complexities that we’ve encountered in Toolbox databases that require technical help in order to import correctly into FLEx:

  1. Some subentries (\se) relate to a specific sense in the entry rather than to the entry as a whole.

  2. Some subentries (\se) have multiple senses.

  3. Some cross references (\cf) or variants (\va) relate to a specific sense in the entry rather than to the entry as a whole.

  4. Some cross references (\cf) or variants (\va) contain multiple items (e.g., \cf run, jog, trot)

  5. Some cross references (\cf) or variants (\va) contain an item that has no target. That is, \cf abcd or \va abcd exists in some entry but \lx abcd does not.

  6. The target of a cross reference or a variant does not always include an \mn field that clearly identifies the entry where it is referenced.

  7. Some \lx fields have content that includes tone marking, while others have content that uses the orthographical convention of not marking tone.

  8. The \ps field is understood to apply to multiple senses in the same entry. (Note: This is normally a problem only if the entry has multiple \ps values but each value is understood to apply to more than one sense, as illustrated below, where senses 1 and 2 are understood to be nouns and senses 3 and 4 are understood to be verbs.)
    \lx
    \ps n
    \sn 1
    \sn 2
    \ps v
    \sn 3
    \sn 4

  9. Complex forms are linked to multiple entries. For example, steamboat is linked to the entry for steam and to the entry for boat.

Successful import of a database that has one or more of the above characteristics will be a challenge for a trained expert and close to impossible for an untrained individual.

Best wishes,

Kevin Warfel
Associate Dictionary & Lexicography Services Coordinator
Rapid Word Collection workshop consultant

7

Thank you Kevin!

Kind regards,

Chris