I have just spent an hour using Safari and bookmarklet to no avail.
Step 1. Log in to online library database to search for articles.
Step 2. Browse through journals and
Step 3. Use PDF bookmarklet to save articles to my database.
Using this means I added about 30 articles to my database.
Great!
But actually, when I go to my database I get 30 versions of the library login page.
The result: I have wasted an hour because Devonthink did not inform me that this would happen.
I know that had I used the Print PDF to Devonthink this would not have happened. But that’s not enough. In this instance the application led me to believe that all was going well. If the PDF bookmarklet doesn’t know how to do the same thing as the Print to Devonthink option, then the application should tell me.
It’s not a bug. It results from the behavior of that Web site, which requires a login to re-access a page. When you invoked that bookmarklet, the URL was passed to a procedure to download the page and convert as PDF. The procedure did its thing, but what was presented to it by the URL was not the page you were viewing, but a login page.
Not all sites requiring a login behave that way. If login permission is automatically recognized, e.g., via a cookie, you would get your PDF as expected. But if the site is set up to require login for each separate access (like my bank’s site) you will get a PDF of the login page.
Now you know.
In such a case, the capture as rich text Service works for selected text/images and the print as PDF capture mode works, as they are working directly from the page content you are viewing. I haven’t checked, but I think the Services capture of selected content as WebArchive would also work on your site.
I understand what you’ve said.
What I’m saying is a bug is that the application doesn’t give you any feedback to indicate that that’s happening. If you’ve told the application to send a PDF to your database, and it can’t do that, because the website won’t let it, then is there not a way for the application to know that and send you a warning?
You see, from the user point of view, he or she has logged on to the website already. And the bookmarklet suggests that it will send the pdf to your database. And it doesn’t.
I know there are reasons why this doesn’t happen, but I think its a pity to spend an hour finding useful documents, and then discover you have none of them. If there had been a warning, you would have changed technique sooner.
So… in brief, the bug I refer to is the lack of warning. I won’t make the mistake again, but maybe someone else will.
I don’t agree that this is a bug. The bookmarklet dutifully took the URL handed off to it and created a PDF file of the page that was pulled by the URL.
You got a PDF of the login page because the site designer deliberately controlled access to the page you wanted to capture. Some sites that require login for access will do that to you, others will not. All highly secure Web sites will do that to you, and most would not accept a second login from a user who is already logged in.
So it’s a good idea to inspect the results of a capture made from a site that controls user access, to make certain that the chosen capture method is allowed. There’s no reasonable way that bookmarklets or scripts could be written in a way to detect all the access control tricks that are used.
If someone told me that a security expert had designed a truly ‘eyes only’ Web site, that prevents printing, clipping or screenshots of the displayed page, I wouldn’t be surprised. I’ll bet that exists, or will soon.
It’s not a bug simply because there is no way for DEVONthink to know that that page requires a login. It’s that simple. Unless the web page designer uses the http login request, where we can detect that a login is required. An in-page login is not such a beast and will fail in your situation. There really isn’t much we can do about it.
Ahum, . I would like to file a COMPLAINT (completely separate frum a Bug, Feature Request, or a Pity). I realize that Christian is busy n’ all having just had a baby and stuff, but when he finds a coupla extra minutes could you please mention that it’s Completely Obvious to me that the next beta needs a feature where I can enter a molecule’s CAS number, and then it should ask me, “Create Materials…? How Many Grams Please?”
After this part, it should then materialize them, like in Downloads or sumthin, and allow me to drag n’ drop 'em out of there, onto my desk.
I just don’t understand how this critical feature has been overlooked before now.
Thank you for your immediate attention to this urgent matter. If you could add this, right away, if not sooner, that would utterly rock, and make me feel really happy inside.
