3 ms·
I was in tears with the amount of manual labor on easily scripted tasks the author spent. Months of work that could have been scripted in a few weeks. You can'
by vorpalhex 6y ago
I was in tears with the amount of manual labor on easily scripted tasks the author spent. Months of work that could have been scripted in a few weeks.
You can't script the human element out, but you can save a whole bunch of time clicking through pdfs looking for forms and filing github issues.
- beckingz 6y agoHow would you go about programmatically determining if a PDF is a form or not?
- prepend 6y agoMechanical Turk? HTTP get each pdf, look for tillable fields? Flag those for review. Even manually, 2 months to review 2000 links is a long time. I don’t think the author meant he literally sat there for 40 days, but likely did other stuff too.
- hoten 6y agoPuppeteer should work. I believe PDF form controls are turned in HTML form elements, which can be queried by Puppeteer.
- abraae 6y agoOur product automates pdf forms in a specific area (HR). From experience, the presence of form fields in the PDF will tell you that it is a form, but the converse is not always true. Plenty of people flatten their pdfs to remove the form fields, or maybe they were never there in the first place. I believe it's sometimes driven by misguided attempts to stop people from "hacking" (aka automating) them. Or pure incompetence. Some other things I would have like to see in the article would be some details of how they handled signatures. We do electronic signatures but it is certainly an area where accessibility comes into play.
- vorpalhex 6y ago+ check for actual form fields + check rectangles as % of form + flag burned pdfs for review
- williamsmj 6y agoI bet the author learned a huge amount of valuable context and domain knowledge by reviewing so many forms "by hand".