bug report 1: with certain pdfs it works perfect, with others not (although DT shows them as PDF+TEXT and I can search in the text, and highlight text, therefore OCR should be fine). And in certain pdfs if I highlight a whole paragraph, in the summary I see only the first words or the first line of the paragraph.
bug report 2: if I click on the link in the summary to jump to the original pdf, in the DT list still the summary file remains the selected one (although I am then in the pdf). If I again click on the summary file nothing happens, I am not able to jump back to the summary file, but I have then to select another file and then again the summary file to return to the summary file itself.
the pdf summary feature is simply great, but the bugs mentioned above (especially that not the full highlighted text is extracted but only part of the highlights) makes it difficult to use. But since we are in a beta, I am absolutely sure that Devontech is able to fix this with one of the next updates… then it will be a fantastic feature.
(although DT shows them as PDF+TEXT and I can search in the text, and highlight text, therefore OCR should be fine)
There is a misunderstanding of what OCR does. OCR does NOT mean the underlying text layer is always presented as you see it in the PDF. It is sometimes the case? Sometimes it can be. Is it always? Definitely not. The underlying text can sometime be reordered in surprising ways. The underlying structure of a PDF is also not what you’re seeing on screen.
People don’t notice this since…
The PDF looks correct. (OCR doesn’t change the appearanace of a document).
People search for words, not paragraphs or sentences.
People rarely export the plain text of a document, where issues can be seen.
Have you tried to copy and paste your highlighted text from ur pdf to a rtf file and check whether you are getting the right text? The issues you have are more likely related to ur pdf than to the function.
Did you use the “back” button to go back from pdf to the summary, it’s always the way DT works and sometimes you need to click a few times (I think).
No, this indeed seems to be a bug: it’s possible to copy the highlighted paragraph (CMD+C) and insert it into a RTF (with CMD+V) without problems. However, the summary tool shows only the first words of the paragraph in the summary-RTF and not the whole highlighted paragraph (as copying to RTF works without problems OCR is absolutely fine). As the documents are with a copyright, I am not allowed to publish them here, but I can provide them to DT developers if needed.
You are right, the back button brings me back to the summary (after first going back in the pdf itself in case I have scrolled there before --> so there is no way to jump back directly to the summary RTF). After jumping from the summary RTF to the PDF with the link in the summary RTF, it is not possible to return to the summary RTF by clicking on the summary RTF. This indeed seems to be a bug.
As soon as these two problems have been fixed, the summary feature will be fantastic and will become one of my most used features!!!
It appears that the summary function is still a work in progress, but wow, this has so much potential for my workflow. I’m chiming in here just to share how my results are different from what I’m reading. When I select a PDF+text file in which I’ve highlighted key headings, no new summary document is created. If it has been, I can’t seem to find it after looking up and down the DT3 path hierarchy.
My use case would be summarizing meeting minutes. The Town Council Secretary underlines main topics, and the previous discussion in this thread leads me to believe the underlines have potential to get picked up as a highlight. This could save me hours of work and allow for quick visual scanning of topics. I imagine creating one long document that serves to chronologically track ordinances, resolutions, and topics of discussion. This makes me giddy.
But this feature has so much potential it’s very exciting! But possibly some bugs to iron out and some interface issues with DTTG. Eg. I edited an RTF in DTTG generated with the summarize highlights function on the mac. All highlights were stripped out and links removed after that edit. Some suggestions for the future – choosing different filetypes e.g. markdown, and removing the pesky spaces and hyphen in the rtf whenever a line-break occurs in the pdf. But very helpful feature.
I agree - this is pretty great. I love the link back to paragraphs personally. My current workflow is to do all my reading/note taking on the Ipad (in PDF Expert), then generate text files of each quote/highlight, which I store in Bookends. But now the incentive to generate a final version of notes in DTP, as I can include the direct link back to the actual page/paragraph. Magical.
I just discovered this Summarizing feature, which works flawlessly for me!!! This is outstanding, and helps me in so many ways in my academic writing.
I think you’re hiding this feature under a bushel - it’s worth illuminating!
Muchas gracias.
