if wishes were fishes ... comment on wishlist ranking
Author: firstname.lastname@example.org (ulrich-merkel)
As long as 200 voters rate it LOWEST, it goes to the top of the wishlist
Let me quote some lines from Ton Blankers Article:
July 14th, 2010 Report on Uniface wish list
On July 9th 2009 the Wishlist application went live. As promised this is the second report on Uniface wishes. Every 6 months the Uniface Lab will publish a report a wish “top 10” and what the plans are to implement wishes in the next Uniface version.
Since the start of the wishlist application in total 138 wishes have been received. The voting mechanism of the Community website has been used to create the following wish “Top 10 July 2010”
- (1,165) Have the possibility to interrupt a Uniface process.
- (-,153) Reduce unnececessary related entity warning (FireSongKT)
- (-,127) Advanced String handling methods (GHAN)
- (2,145) Support for Regular expressions (Sochaz)
- (9, 124) Grid dynamic column resize/hide/show/move (addice)
- (4, 124) Code sensitive in UDE (apicaud)3.
- ( 3, 116) Open Hitlist / Pagination (GHAN)
- (5,112) Color Management in uniface 9.x (gypsilon)
- (-,108) Ability to use .net forms)
- (8,108) Goto definition for procedure calls and operations (ishankter)
Between brackets previous position and wish rating=number of votes * average rating
|Jul 16, 2010 8:54 AM|
|Hi Iain, good point that we need a better participation of the members; but when the number of voters is +/- even for each wish, a simple average will do the job. What is rated as HIGH will be on top of the list, what is rated LOW is at the end. Success, Uli|
|Jul 15, 2010 4:15 PM|
|Where is the schedule for the future. I need for 2 customers Oracle 11gR2 on HPK and RSD but I can't inform my customers about the support on this database. Also, can you specify in the PAM, it's the Oracle client 32 bits only on 64 bits plate forme.|
|Jul 15, 2010 9:47 AM|
|I agree, besides, average * no of voters = total points assigned. However, they need to reflect the popularity of a suggestion in the totals. All I can suggest is that we all try to do what I try to do, which is to rate everything that is posted, rather than just for or against specific interests. That way the baseline is more fixed. Unless you have a better statistical method for prioritising from variable inputs?|
|Jul 14, 2010 5:33 PM|
|As I mentioned before, using the calculation "rating=number of votes * average rating" puts´the 1 of 100 voters (=100) on top of the list, while a 10 from 9 voters = 90 comes last. is this a fair calculation? I don't think so. Very disappointing, Uli|
A quote from my discussion paper:
current wishlist statistics (100113)
it took me some 100 mins to extract the rating and no-of-voters from all the detail screens
(others have the ability to export wish-data into an excel format)
But here is a first little statistic:
||rating 100113||voters 100113||rate*voters|
As we see, only some 7 out of 950 members vote this makes less than 1% !!!
Should we really base our selection to such a micro-minority of the registered members
which wishes makes it to the Top10 and get viewed/fulfilled, ..
Assume there are 10 registered employees of a softwarehouse
rating all "their" wishes to "10"
and all "foreign" wishes to "1"
what about minorities or softwarehouses with only 5 registered emplyees?
|Jan 21, 2010 12:23 PM|
|Well spoken Uli /Sven Hedqvist|