Light on the horizon...
Lukas, thank you for the icons, they are now in the repository. I've built the app with the new icons. I'm now converting the documentation to inDesign which cooks up better pdfs and html than MS Word. Well at any rate much lighter html than word did. I think it's worth it for smaller apps and prettier pdfs.

When this is done and I cook up new apps from scratch I will then

If there haven't been many entries, it's good news...
We're down to the small details of the interface:
  • Should get the code definition be separate: one for selecting text, one for clicking on lists...
  • Should I redo import or new pdf so that it opens a file witha  pdf attached
  • Should I add an other button onto the new project panel (I think so!)

But mostly it's down to issues with the icon; how reductionist is too reductionist? What if icons aren't visually distinct? I know I have my attachments, so it's hard working with our artist, but it's also informative, and wonderful (and painful) to see tams grow beyond my personal project.

Biggest headache? Working with subversion for the documentation. Maybe more on this next time. 

Interface change
I realize how TA violates HIG issues in favor of efficiency of the interface in a lot of ways. I rearranged the interface a little. It was just bugging me; the arrangement didn't make visual sense to me. Not in terms of utility but visual balance. Now the "Apply code" button is a unified element across all document types.

Two critical bug fixes, one which should be back applied to TA3. But I think I'm just going to just plough ahead with TA4.

Well, I shouldn't have
4 people are (in theory) beta testing; but I've already made tons of changes; patched some bugs, or at least moved the bugs around.  They're still there. Just crashed, but have no clue why. Yes, I know where it crashed, but the statements seem totally benign!

Meanwhile have started redoing the documentation. I'm coding my own data. AND TA4 is feature complete! There may be small tweaks to the interface, but as far as I'm concerned it has all the goodies that users need to get going!

Shouldn't release too soon
Things are bopping along. I've actually used TA4 a lot to code zines that I've collected on the street medic project. That has helped to identify key issues, like the need for a good search mechanism. I've implemented a crude one.

I keep tripping across un-repeatable bugs too; so that's another reason to not release early.

Alpha version is complete!
With 48 commits; first commit was May 19; last was May 31!!!! That's really amazing to have added the new features (with a lot of sweat, really a couple 100 hours), work I thought would take all summer. I'll make a release soon, once I make some quick documentation.

Dancing and coding...
Yes I've danced a lot this weekend...

But I've also coded. A lot.

At this point TA4 is about 90% done. Here is the state of the repository:

Searching for all file types: done

Coding all file types: done

Select and scroll to parts of a pdf: done

I have also dramatically simplified the document windows and brought them up to date with pdf/image windows.

Still to do, The results media player needs to show selections for pdf

split views need to be saved for image and pdf windows

image windows need to broadcast change

PDF Coding
Well, pdf coding is half finished. It was a brutal day. Dropped the working class from the lab setting into tams analyzer and boom, nothing worked. Spent the day chasing bugs.

So, I think (but am probably wrong) that the hard half of coding is done. The other half, text coding in pdfs, should be handled  by Apple's pdfkit through the class pdfselection. I've developed the scheme for encoding information, not I just need to code it.

But it's northwest folklife festival, and I'm taking it easy this weekend. balance and swing!

On the art of selecting rectangles in pdfs
To start with, coding of images is complete. Yup, it's a done deal. All it needs is an icon.

On to pdfs. 2 days and a lot of headaches, several failed attempts at creating code, I have succeeded in creating a selection rectangle that works with PDFViews. My knowledge of programming graphics is minimal. I was highly dependent on the code that Apple circulates called Crop Marker or cropping view or some such. It took a lot of mods, however to get it to work. Here are some breadcrumbs to help other fools who go down this road:

Also, I built a new release of TAMS 3 with the fixes to rtfd reading (as well as older bugs). Available as always at the tams website

1. you need to override drawpage not drawrect

2. you need to do a double coordinate transformation: global->pdfview->pdfpage to get your selection to go to the right spot. This rect sizes correctly if you scale.

3. you need to check the page whenever the mouse goes down and you draw, otherwise your selection rect appears on every page of the pdf.

So this means that the biggest stumbling block to coding PDFs has been crossed. There's a lot of work to be done, but it's quite doable now that the basic mechanism for marking pdfs has been worked out.

TAMS will support both pdf's native "text" selection, and rectangular area image selection.

Only one mystery remains: what method initializes my subclass of PDFView?

Fixes & Progress
Fixes: Got a note from someone who found that multiuser tams didn't work with rtfd's--after investigation found out that rtfd's are bundles and have to be handled separately from everything else. Oy.

Progress: result windows show the image but don't select the crop mark.--something to do tomorrow along with fix prev and next to use the new version of jump to row.


Log in

No account? Create an account