|
Author |
Topic: Board code is overzealously rejecting non-URLs (Read 1887 times) |
|
still_studying
Guest
|
 |
Board code is overzealously rejecting non-URLs
« on: Jun 29th, 2007, 3:12pm » |
Quote Modify
Remove
|
I had to resubmit a posting four times before I finally discovered that it was being rejected for having the "URL" of "i dot e dot". This wouldn't be so much of a problem, except that (1) the board clears out everything that has been typed when you hit the "back" button, and (2) there is no way to determine what it is spuriously identifying as a URL, except through repeated trial and error. (Also, (3) greate pumpkin, this server is slow today!) Suggestion: instead of a straight "no URLs, hit the back button (and lose all of your typing)" screen, how about putting the filled-out form on the error page so that it can be corrected and resubmitted? I'll even do the coding if whoever runs the site will get in touch. (Due to the holiday, it might take a few days for me to reply.) (I assume the code is all Perl?)
|
|
IP Logged |
|
|
|
JimIvey
Moderator Senior Member
    
Posts: 2584
|
 |
Re: Board code is overzealously rejecting non-URLs
« Reply #1 on: Jun 29th, 2007, 4:18pm » |
Quote Modify
|
First, my apologies for not fixing the issue in the manner you describe but instead offering "why don't you"s. I'm moderator solely for the purpose of spam killing. The real moderator would have to make those mods. "why don't you" #1: Why don't you use "ie." instead? "why don't you" #2: Why don't you register? You've posted enough as a guest that you would be permitted to post URLs by now if you had registered. "why don't you" #3: Why don't you use FireFox? When I go "back", the form is still filled in. We had a long dialogue about that a year or so ago -- some browsers keep your form information when going "back"; others don't. Again, I apologize for suggesting you fix your own problem (as it's one that could be shared by others and is a legitimate problem), but that's all I can do at the moment. Regards.
|
|
IP Logged |
-- James D. Ivey Law Offices of James D. Ivey http://www.iveylaw.com
|
|
|
still_studying
Guest
|
 |
Re: Board code is overzealously rejecting non-URLs
« Reply #2 on: Jun 29th, 2007, 6:21pm » |
Quote Modify
Remove
|
on Jun 29th, 2007, 4:18pm, JimIvey wrote:First, my apologies for not fixing the issue in the manner you describe but instead offering "why don't you"s. I'm moderator solely for the purpose of spam killing. The real moderator would have to make those mods. |
| Hey, I'd be happy to help. I even mentioned it in my original post. I'm not half bad(*) at Perl coding. (* - Understatement of the week; well, ok, the hour.) Quote:"why don't you" #1: Why don't you use "ie." instead? |
| Because that would be improperly punctuated! Quote:"why don't you" #2: Why don't you register? You've posted enough as a guest that you would be permitted to post URLs by now if you had registered. |
| I know, I know. I should. I really should. I tried once. I don't know why I couldn't get it to work. I'm such a dork. Quote:"why don't you" #3: Why don't you use FireFox? When I go "back", the form is still filled in. We had a long dialogue about that a year or so ago -- some browsers keep your form information when going "back"; others don't. |
| This is NOT my choice; unfortunately, I cannot get an internet connection from home due to technical reasons, so I am stuck using the software provided on whatever computer I borrow to read/post. Also unfortunately, the cable provider at this place has some proprietary software which prevents me from bringing and connecting my own machine. Quote:Again, I apologize for suggesting you fix your own problem (as it's one that could be shared by others and is a legitimate problem), but that's all I can do at the moment. Regards. |
| It's quite all right. The only minor issue I have is that I really would be more than willing to "fix (my) own problem" by helping out on the board's code -- thereby also helping others who are stuck using (or, due to some inexplicable defect, choose to use) IE -- and I really don't mean to attempt to demand the imposition of a burden on the host of this fine establishment. And there really are others running into the same thing; I read a post today that mentioned it (somewhere in the posts made since 6/22). Glad you liked tesseract, BTW. Personally, I find the accuracy too low for most purposes, but at least it's a lot better than retyping whole documents.
|
|
IP Logged |
|
|
|
JimIvey
Moderator Senior Member
    
Posts: 2584
|
 |
Re: Board code is overzealously rejecting non-URLs
« Reply #3 on: Jun 30th, 2007, 9:57am » |
Quote Modify
|
on Jun 29th, 2007, 6:21pm, still_studying wrote:Hey, I'd be happy to help. I even mentioned it in my original post. I'm not half bad(*) at Perl coding. (* - Understatement of the week; well, ok, the hour.) .... I know, I know. I should. I really should. I tried once. I don't know why I couldn't get it to work. I'm such a dork. |
| Let's see.... trust the perl source to someone who can't register.... ? ;-) Just kidding -- but you recognize in the incongruity, right? on Jun 29th, 2007, 6:21pm, still_studying wrote:This is NOT my choice; unfortunately, I cannot get an internet connection from home due to technical reasons, so I am stuck using the software provided on whatever computer I borrow to read/post. Also unfortunately, the cable provider at this place has some proprietary software which prevents me from bringing and connecting my own machine. |
| All that is necessary for evil to prevail is for good (wo)men to stand by and do nothing.... Take your computing infrastructure into your own hands! http://librenix.com/?inode=3745 And, if you absolutely must use computing's least common denominator (or rebooting under a different OS renders the access inoperative).... http://u3.com/ on Jun 29th, 2007, 6:21pm, still_studying wrote: Glad you liked tesseract, BTW. Personally, I find the accuracy too low for most purposes, but at least it's a lot better than retyping whole documents. |
| It's great! I'm hatching plans to set up a cron job to OCR all my received faxes to make them text-searchable. Something like this: Quote:find . -iname '*.tif' -exec tesseract "{}" "{}.txt." \; |
| Then, if I had a fax stored at 2007-01-01_Fax.tif, the text "equivalent" would be at 2007-01-01_Fax.tif.txt. Finding the searched text there would direct me (by reversing the file name translation -- i.e., ignoring the .txt suffix) to the fax image. Of course, for efficiency reasons, I might limit it to faxes received in the last week or so and/or exclude files for which a text file exists and is newer than the fax. How cool is that?!?! So what if the fax header and other extraneous junk gets included in the text file? The only problem would be if the searched-for text was not correctly recognized by tesseract, but accuracy seems to be about 98% or better -- what are the odds that the particular text I'm looking for (likely to be repeated in multiple documents, such as a foreign application number) would be incorrectly recognized in all instances? Acceptably low -- 98+% chance of finding the information vs. 0% the old way -- significant improvement in my opinion. Thanks again. You might write the real moderator directly if you'd like to take a chance at making your mod. Regards.
|
|
IP Logged |
-- James D. Ivey Law Offices of James D. Ivey http://www.iveylaw.com
|
|
|
still_studying
Guest
|
 |
Re: Board code is overzealously rejecting non-URLs
« Reply #4 on: Jul 4th, 2007, 3:55pm » |
Quote Modify
Remove
|
on Jun 30th, 2007, 9:57am, JimIvey wrote: Let's see.... trust the perl source to someone who can't register.... ? ;-) Just kidding -- but you recognize in the incongruity, right? |
| Grrr. ;D Thing is, it's basically a cut-and-paste from the code for a "preview" page. If you hit "preview", you get a form again. If the site thinks you've typed a URL (even if it is bizarrely mistaken) you (well, ok, *I*) lose what's been typed. Oh well, no matter. I'll just be careful to cut and repaste the text so that I'm sure it's saved in the clipboard, if nowhere else.
|
|
IP Logged |
|
|
|
|
|