Selecting "On the Fly" which links to follow & download
|jmazor||06/15/2004 10:55 pm|
|Oleg, you obviously recognize that it is very important for the user to be able to SELECT which links get followed and downloaded and which do not. I THINK I HAVE A REALLY GOOD IDEA HERE. `LOVE TO KNOW WHAT YOU THINK.
Design Challenge: But even your very creative URL and content filtering tools have to deal with the very difficult problem of getting the user to ANTICIPATE what links may be present and DESIGN rules (i.e., filters) for the program to use as it flys through the websites that the user would make if the user was making the filtering selections one at a time.
Idea: Let OEP operate just as it does now, but have it "step through" the download process, stopping at each new webpage permitted by the filters and asking the user "Download this page? Y/N". That way, the user can filter as tightly or loosely as he wants, then see each resulting page, and then decide whether to keep the page.
Example: You set the Level Limit to 2. OEP downloads the Homepage, which has 3 links ("A", "B" and "C") that are permitted by whatever filters you set up. Then OEL follows "Link A" to its Level 1 page, displays the page and the prompt, "Download this page? Y/N". If user answers "No", OEL looks back at the homepage and follows "Link B" and displays its page and the prompte, "Download this page? Y/N". If "Yes", OEL next displays a linked Level 2 page and presents the "Download?" prompt, etc.
Result: User can be much looser about setting up filters; results will be exactly what the user wants; no time spent designing and redesigning filters; no time spent re-running downloads.
What do you think?
|Oleg Chernavin||06/16/2004 04:26 am|
|There is a problem - if a page has dozens of links (a typical Web page), it will be a challenge for a user to answer every question. A much easier solution is ti start loading a Project and hit F9 key, wait for the first page to be loaded and then go to the Queue tab and abort unwanted links. Then press F9 again to follow links and watch the queue. When you see a link that you don`t want, press F9 again, abort and go further, etc.
We have some ideas on how to make this process completely easy and convenient for users, but it will take time to design it.
|jmazor||06/16/2004 05:35 pm|
+But I`m referring ONLY to links to other web pages (maybe a couple of dozen), not the scores of references to gif, jpg and other files that are used on the home page.
+The list in the Queue tab shows potentially hundreds of URL`s because it includes the URL`s for each file used by the webpage in addition to the URL for each actual link.
+Also, how`s a user to decide whether to accept or reject a webpage just based on the URL?
+Wouldn`t it be much quicker and easier for the user to see a webpage and quickly click Y/N than to study a long list of URL`s in the queue tab list and guess whether he wants each one from just the name of the URL itself?
+One other possibility that may be easier to program: redirect the Queue list to a screen with a checkbox next to each URL. Then the user could tell the program which URL`s to keep in the Que and use, and which to ignore.
> There is a problem - if a page has dozens of links (a typical Web page), it will be a challenge for a user to answer every question. A much easier solution is ti start loading a Project and hit F9 key, wait for the first page to be loaded and then go to the Queue tab and abort unwanted links. Then press F9 again to follow links and watch the queue. When you see a link that you don`t want, press F9 again, abort and go further, etc.
> We have some ideas on how to make this process completely easy and convenient for users, but it will take time to design it.
> Best regards,
> Oleg Chernavin
> MP Staff
|Oleg Chernavin||06/17/2004 04:21 am|
|> +One other possibility that may be easier to program: redirect the Queue list to a screen with a checkbox next to each URL. Then the user could tell the program which URL`s to keep in the Que and use, and which to ignore.
Yes, I like this more. Do you mean that OE should display a list of links with their natural names and URLs?
Also, what should happen when OE finishes loading 10 approved links from the first page and gets new links from them? Should OE show 10 screens sequentally to continue the process?
What about another solution - a user would simply drag the desired links on a page to a folder and all links in dragged selections will be loaded? This would work well for small sites, while complex sites will still require the old-fashioned adjustment.
|jmazor||06/17/2004 01:15 pm|
This is rough but I hope useful. When I look at the “Map” generated by a download, I see something already pretty close to what I’m thinking.
Step 1 - OEP makes a Pick List of the Webpages Available for Download: OEP runs, but at first does not actually download files. Instead it display a TREE & BRANCH type of Map-like “Pick List” of all the webpages (showing the Natural Name and the URL) that it would download in accord with the Level Limit, URL filters and Content filters. However, there would be a checkbox in front of each item. The list would not contain any reference to the jpg, gif or other files that usually appear in a Map tab.
Step 2 - User Selects. The user checks to indicate which webpages he wants downloaded. Example: He might select two out of however many level one sites linked to the home page. He then might look at the first selected level one webpage and decide that he does not need any of its level two pages. On the other hand, for the second level one webpage, he might select some but not all of the level two pages listed within it.
Step 3 - OEP Downloads the Selected Webpages. OEP downloads the selected webpages and their components, skipping the webpages that the user did not select.
What do you think?
|Oleg Chernavin||06/18/2004 05:15 am|
|Yes, this can be done. I will work on it.
|jmazor||06/18/2004 08:34 am|
FANTASTIC. THANK YOU SO MUCH.
|Oleg Chernavin||06/18/2004 08:50 am|
|However this will take time. I am not sure that it will be implemented by 3.3 or 3.4 version. Maybe later.