Outclass sucks

[Joe Doe]

Hey there,
great idea to provide a tool that works with outlook (outclass, that
is). I thought it would do what I need, but it doesn't. I actually hate
that thing but I can't find any alternative. Here's why:

1. bloated - takes forever to lead
2. design flawed - creates its own buckets instead of adapting to
popfile

Disadvantages:
1. If you already have a decent corpus with popfile, but are looking for
an easier way to reclassify instead of copy/pasting the url, you cannot
use it with outclass - you will have to retrain popfile from scratch.

2. If you want to tie in to a remote popfile "server", you can't do that

3. If you have a Mac and a PC, and want to use popfile on both, using
outclass will make you lose all the trainings you have done once you are
back on the Mac, since outclass doesn't train the exisiting popfile
corpuses.

Actually, this thing pretty much sux. You can only use it if you use one
PC, and no other machine ever. All your email will have to go through
that one PC. You lose tons of flexibility there.

Now that I complained (rightfully so), I also want to offer a solution.

Make outclass interact with popfile instead of "using" it's bayesian
capabilities to do your own proprietary thing. Make outclass so that it
figures out from the email which url needs to be fetched and
reclassified to train popfile itself instead of doing your own thing.
Then it will be truly flexible and a kickass product.

How come noone has thought of this before?

 

[Milly Staples [MVP - Outlook]]

Perhaps because you are complaining to the wrong people? Try posting your
suggestions to the programers of the application - it is not Microsoft.

--
Milly Staples [MVP - Outlook]

Post all replies to the group to keep the discussion intact. Due to
the (insert latest virus name here) virus, all mail sent to my personal
account will be deleted without reading.

After searching google.groups.com and finding no answer, Joe Doe asked:

| Hey there,
| great idea to provide a tool that works with outlook (outclass, that
| is). I thought it would do what I need, but it doesn't. I actually
| hate that thing but I can't find any alternative. Here's why:
|
| 1. bloated - takes forever to lead
| 2. design flawed - creates its own buckets instead of adapting to
| popfile
|
| Disadvantages:
| 1. If you already have a decent corpus with popfile, but are looking
| for an easier way to reclassify instead of copy/pasting the url, you
| cannot use it with outclass - you will have to retrain popfile from
| scratch.
|
| 2. If you want to tie in to a remote popfile "server", you can't do
| that
|
| 3. If you have a Mac and a PC, and want to use popfile on both, using
| outclass will make you lose all the trainings you have done once you
| are back on the Mac, since outclass doesn't train the exisiting
| popfile corpuses.
|
| Actually, this thing pretty much sux. You can only use it if you use
| one PC, and no other machine ever. All your email will have to go
| through that one PC. You lose tons of flexibility there.
|
| Now that I complained (rightfully so), I also want to offer a
| solution.
|
| Make outclass interact with popfile instead of "using" it's bayesian
| capabilities to do your own proprietary thing. Make outclass so that
| it figures out from the email which url needs to be fetched and
| reclassified to train popfile itself instead of doing your own thing.
| Then it will be truly flexible and a kickass product.
|
| How come noone has thought of this before?