MMLT Reader Prototype System

[akhtet]

I just quickly coded a prototype to demonstrate my new project idea.

http://reader.mmlt.org/reader.py
http://akhtet.mmlt.org/cgi-bin/reader.py
http://idro.mmlt.org/cgi-bin/reader.py

Please let me know what you think.
1. Should this idea be adopted and developed further ?
2. If so, what kind of features should we provide ?

I am looking for people who wants to join me in developing a real version of this system.

[akhtet]

Ma HHH suggested showing Wiki Myanmar updates on the feed system

[victor_lin21]

Great! Here are features request.

- Tags to Use Tag Cloud Effect.
- Nice tamplet parser for each post, mean each post shall parse sudden amount of text aka say first 600 words count and after that read more to link back source.
- The parsed post display shall be very nice "dialog text box" style, probably css+image design. See the following for reference:

http://planet.gnome.org/ (Very close to this)
http://planet.fedoraproject.org/
http://en.wikipedia.org/wiki/Tag_cloud

Ummmmm, more later..

[akhtet]

Great! Vic also come up with some good references

I'm thinking to categorize by 3 aspects

1) By tag (by information content)
Posts are automatically tagged by the contents.
These are feeds, not permanent bookmarks, so I see no point of manually tagging it either.

2) By source (mmgeeks? mmitpro? etc.)

3) By author (ThetThwe? victorlin_21? )
That means even if you write in multiple sites, if you still stick to same user name, all your posts shall be under your name.

Tag Cloud: I'm not sure now useful it would be in our case because the tags you see are automatically generated, not user generated.

[Thet Twe]

Lu soe, you spelled wrongly to my name.

Humm, I think you misunderstood about cloud tags. Cloud tags are auto generated as well. Cloud tags are going by user hits. If user clicks more on this topic, font will be bigger in cloud tags. Cloud is like the suggestion section to get the people attention. I think it's good idea to add.

If you want to add tags by author, you should separate with Normal Tags. I think. Because combing looks hard to find when there are so many tags.

Apart from tags, you should display site by site also.

E.g MMGeeks - Burmese Computing Network

Under - Latest Topic feeds will display.

E.g Bashroan Blogs

Under - Latest Topic Feeds will display.

Anyway, it's great job ! Keep it up bro.

[akhtet]

MMLT Reader Prototype 0.2 is up.

New features:
1) Separated tags and authors (Support author names in Unicode).
2) Auto-classification of posts into human-defined categories (i.e. tags).

[victor_lin21]

Cheee...

Tag Cloud
http://webdesign.about.com/od/csstutori ... 011407.htm

Code: Select all

If TAGNAME > post.count(100)
then
parse.html<a href="" style="font-size:100pt">TAGNAME</a>
elseif TAGNAME < post.count(100) && TAGNAME > post.count(50)
parse.html<a href="" style="font-size:50pt">TAGNAME</a>
elseif
........
'''
Well, use nice css for styling the font, of course
'''

Round corner or Ballon dialog box, just roll with some nice css, such as;
http://www.toolsoup.com/extra/balloon.html
Right Click > View Source for css and html usage.

Actually, what I'm trying to say is something like this. Ask TT for designing CSS shape like that kind of dialog ballon with graphic. He can do it in 1 second.
A few resources are here: http://www.kholix.com/wiki/index.php/Css#Round_Corner

[akhtet]

I'm fine with tag-clouds, but that is still very advanced for the current prototype.
Tag clouds well be added sooner or later.

I am more concerned about essential functions.
The idea is basically half a search-engine and half a bookmarker
I'm also thinking to allow user account creation and customization.

To continue for alpha version, I am going to need the following modules on MMLT server.

1) Universal Feed Parser http://feedparser.org

2) Django + ModPython

3) MySQL database + Python interface
(or) pySQLlite (it's built-in if you have python 2.5, i think)

[akhtet]

Quickly updated to version 0.21.
The author names retrieved from the feeds have been normalized.

[aTxIvG4001]

Quickly updated to version 0.21.
The author names retrieved from the feeds have been normalized.

author names doesnot allow space? why?