Note - I have returned to Tinderbox after a couple of years, so got rustly with simple stuff like keyboard shortcuts.
I am inspired by LLM models and am wondering this time around I can learn Tinderbox faster. So I am using a LLM model to chat with the native help manual provided by Eastgate. But it is not giving me enough answers.
Would you guys know of any downloadable Tinderbox tutorial other that the one provided by Eastgate?
Just an image of me chatting with the Tinderbox manual below.
I heartily endorse what @archurhh says about The Tinderbox Way. I’m now on my third or fourth reading of this, and am finding new material on almost every page. It’s valuable not only as a learning primer but also as a source of content whose application isn’t apparent on first reading.
Looking at the screengrabs, if you are trying to find out about shortcuts look at Reverse Look-up Map. By the way the answer to the tab pane focus shift is [Tab]+[Opt] which toggles input focus: $Text → first Displayed Attribute (if any) → main view → $Text.
I think( from the screen grab) the last version of PDF of app Help was the v9.2.0 version you have. I say that as I made it: it is made in a similar way to the aTbRef PDF but requires access to the source TBX file that makes Help.
Also in the TB app, and opened via the Help menu are two further help/training PDFs “Getting started in Tinderbox” and “Actions and Dashboards”.
Over the last couple of years we’ve/I’ve been creating and curating a ton of videos that they may find useful: Mastering Tinderbox: Training Videos (Complete List). If you have a specific project ide, please create a thread about it and we can guide you through it.
Or you can print the TBX to PDF - see note (in the TBX, not on the website) ‘print-export-source notes’. note that unless you can purge an old PDF from the LLM, you might want to hold off on the latter as v9.6.0 should be with us soon. I just checked and there are 109 new/changed pages + over 60 changes/fixes not explicitly marked but summarised in the change log.
Your approach is quite interesting. When ChatGPT was first out, I asked it about Tinderbox and was unimpressed as most answers seemed to come for Adobe [sic] Creative Suite documentation. Clearly using a closely curated LLM, albeit smaller, means the LLM isn’t polluted with irrelevant material: asking about Tinderbox doesn’t involve knowing all of Wikipedia, for instance, nor about other software in general.
Is there an article/link explaining how you make such a private LLM model? Is it free or only paid use?
I’d be interested to know if there are questions to which you can’t get answers. Why? The problem might be too small a model. But, sometimes I see issues where the way in which the user phrases their question or the terminology used is the impediment, and this helps point people to the right information—or add new articles if needed.
Yes please! The hardest part of writing a public resource is knowing where it unintentionally fails for the reader.
It will also be useful insight into the (in)ability of ChatGPT to parse information to the correct insight. A new skill for us all with be writing in a way that assists ‘AI’ understanding—given that code has no literal semantic understanding of written/spoken language. How that is to be done is a current journey of discovery. But a necessary one if people are—as it seems—going to migrate to using ‘chat’ type interaction, text or speech, with search resources.