RowTags - what are they good for?

So we’ve implemented rowTags on fields (mostly to be used on listbox fields). They allow you to assign data to a row without it being visible. It travels with the row and it’s saved with the stack, so no real thinking about them once they’re set.

set the rowTag of line 1 of field <field> to <value>
get the rowTag of line 1 of field <field>

There are some subtle differences that make a big difference to how rowTags are copied from one field to another. If you say

put line 1 of field <field1> into line 1 of field <field2>

the rowTag will be copied, as you are copying the line. If you say

put the text of line 1 of field <field1> into line 1 of field <field2>

the rowTag will not be copied, as you are copying the text not the whole line.

1 Like

Just FYI, there’s already somethings like that in form of ‘the linkText’
set the linkText of <line|chunk> to
And similarly you can set ‘the metaData’ of …
set the metadata of line 3 field “description” to “extra information”

I did not know that.

You can set the metadata also on a chunk:

Set the metadata of char 1 of line 2 of field 1 to “XXX”

If you add rowTag it would be nice to make the styledText array aware of rowTag if it is not automatically added.

Also HTMLText and rtfText.

Currently I do not see what rowTag adds that metadata does not provide.

I wasn’t aware of the metadata property when I did up the rowTag. I always thought having a rowTag would be something good to have, and apparently it is as it’s already part of the engine (just with a different name).

2 Likes

I guess this feature will be binned as it’s already in the engine (so much stuff in there, how do you even know what’s in there?!)

1 Like

Ask. A lot of us have been using this engine since last century. We have little to show for the gray hair except the odd moment to help an xTalk dev.

3 Likes

I’ll ask in the future :slight_smile:

2 Likes

It is also very rewarding to use Claude as much for coding as for documentation and checking on ideas and checking of the truth of code against the dictionary and the language guide and all the other fabricated sources. LLMs write doc files and feature lists so effortlessly. But I know they need guidance in this, all an endless task. Just saying. We will see how all this plays out in a few years from now.

I respect your time. I, and the others here, would do our best to share anything we’ve learned which will save you time.

Ease is useful, accuracy more so.

Just yesterday the Network Solutions chat bot insisted their control panel had a feature it doesn’t have. Eventually it relented, as LLMs do.

The most proficient use of any system which enchantingly emulates language but can’t understand a word it types, is in the hands of a user who has already acquired expertise. They’re good typists, a step up from the hundredth monkey.

LLMs which compare and match written code, existing documentation and good questions by creative but unstructured human brains (see our discussion threads here) are extremely helpful in systematic endeavors, hallucinations are minimal in such a setup. But please stay with what you have, as Emily says “I am not your mother.” :face_blowing_a_kiss:

1 Like

What we have is a typing assistant and a wealth of human knowledge to refine it. Seems a useful combo, consistent with other industry results.

I’ve basically spent (coming up on) 5 years studying and tweaking the IDE stacks and libraries and docs, so hopefully all of us chiming in can help prevent anyone from wasted time (or tokens) reinventing wheels (although sometimes I think reinventing can be good too).

1 Like