[Mailman-Developers] Boilerplate and content filtering [was: Introduction and Project Discussion]

Sreyanth sreyanth at gmail.com
Mon Apr 15 09:35:12 CEST 2013


On Mon, Apr 15, 2013 at 10:01 AM, Stephen J. Turnbull <stephen at xemacs.org>wrote:

> Sreyanth writes:
>
>  > Also, I would like to hear more about : Boilerplate stripper AND Better
>  > content-filtering / handling error messages.
>  > ​Boilerplate stripping is trivial to understand. But, can anyone
> elaborate
>  > on Better content-filtering / handling error messages?
>
> But boilerplate stripping is not necessarily trivial to implement,
> because it's not always clear what boilerplate is.  I think it might
> be a good idea to save it off and provide a link rather than discard
> it, which leads to interesting questions of storage, shared links for
> true boilerplate (storage compression of repeatedly encountered text,
> yes, but more important the link will turn purple so you don't need to
> click on it in the next message from that user!), and user interface
> in general.
>
> Yep! But, how about this?
Just hide the boilerplate in the email, giving a link to click. When
clicked, use js to unhide the boilerplate. This would not anyhow require
separate storage. Suggest me something if this is bad!
​​

> Content filtering is mostly going to be about MIME handling: choice of
> the appropriate text/* part and things like that, removing
> images/video/etc where the list prohibits them, converting HTML/
> wordprocessor attachments to plain text, removing MIME parts whose
> Content-Type doesn't match filename or perhaps file(1) magic in the
> content, etc.
>

This is cool! I have done converting the HTML attachments to plain text in
one of my projects, but never worked on wordprocessor attachments (some
library for Python should be there, will check!). So, for this, we will
additionally have to provide the admin with various other options, like to
prohibit images, videos etc. Instead an additional option may also be given
to the admin like using the boilerplate idea you proposed. Store them
somewhere, display these as attachments to the email.​​ Correct me if I am
in the wrong path!


> I can also imagine content filtering (or scoring!) based on word
> choice ("WTF" OK, spelling it out not :-).  Also content filtering
> based on stripping out the quoted from top-posts and replacing them
> with links (after checking that the quoted material is indeed
> available in the archive!)  All coming with on/off options, at least
> for those who remember the IBM 360saurus and other dinosaurs and still
> prefer mail to web. :-)
>

So, we will have to index the archives properly so that even if the post is
not quoted properly in the email, it must be linked to the appropriate
material. This will be cool, isnt it?​​



> Error messages (I think this means delivery status notifications (DSN)
> from mail servers) are a similar kind of problem to text-based
> filtering, though somewhat more stylized.
>
> ​Okay! So, these error messages are to be notified to the sender in a more
user understandable fashion. This is what you meant right?​



> Steve
>



-- 
*Yours Sincerely*
*
*
*Mora Sreyantha Chary*
*Computer Engineering '14*
*National Institute of Technology Karnataka*
*Surathkal, India 575 025*


More information about the Mailman-Developers mailing list