CellarTracker Main Site
Register for Forum | Login | My Profile | Member List | Search

v 3.0.1 release notes

 
View related threads: (in this forum | in all forums)

Logged in as: Guest
Users viewing this topic: none
  Printable Version
All Forums >> [Cellar Talk] >> Release Notes >> v 3.0.1 release notes Page: [1]
Login
Message << Older Topic   Newer Topic >>
v 3.0.1 release notes - 4/5/2005 10:18:41 PM   
Eric

 

Posts: 5743
Joined: 10/10/2003
From: Seattle, WA
Status: offline
DRINKABILITY REPORT
For a number of users, the drinkability report is their most essential tool on the site. I made a number of changes and enhancements to make this view more useful and actionable for folks:
  • The report used to calculate its drinkability score assuming purely linear aging. Now I have fleshed this out considerably with a number of different algorithms including, bell curves that are early, late and centered; a fast-aging curving; a twin-peak curve for wines that are accessible both early and late with 'sleepy' periods in between. You can pick a specific algorithm if you prefer it, but the default drinkability score actually applies all of the different algorithms based on specific attributes of the wines. I will post separately on this.
  • The drinkability score used to take into account all consumed bottles, even those sold, traded, returned etc. Now there is an option to only include bottles that you have drunk (or which you have marked as missing or consumed by a family member).
  • The report used to treat all bottles as equal, regardless of bottle size. Now the report calculates based on 750ml equivalent. So if you open a 375ml bottle, that only counts as opening half a bottle. The actual count of bottles (if different) is still available by hovering over the various numbers.
  • Now the report breaks out pending bottles from those in stock, although all bottles are taken into account for calculating a score.
  • The report can now be filtered by bottle size, storage location, or bin. The total number of bottles for the wine are still shown (as the drinkability score is computed on the basis of all bottles), and there is an additional indication showing the number of bottles for that specific bin, locale or bottle size.
  • NEW, LATE-BREAKING CHANGE! I was in the midst of rewriting the document that describes the availability report in detail, and I realized that the algorithm has been missing one of Paul Homchick's original ideas to help call out wines past their window. In his original algorithm, he added an additional 100 points to the drinkability score for any wine past the set drinking window. This was actually very, very easy to add so I added it.

You can read more here on the upgraded drinkability report.

MISCELLANEOUS QUERY IMPROVEMENTS
The grouped wine summary can now be filtered by keyword, wine name, wine locale, bottle size, storage location, storage bin and purchase price. The bulk-add valuation view can now be filtered by wine name, wine locale, and bottle size.

COOKIE HANDLING
I changed a number of things about the way that the site handles cookies. Deletion is handled more cleanly now by setting the cookie expiration to a date in the past. Also, all of the cookie deletion is now handled centrally in one ASP page. There use to be problems that would occur if a user logged into http://www.cellartracker.com and then later logged into http://cellartracker.com (notice the missing www subdomain) or vice versa. This is now handled by setting the domain property of the cookie to "cellartracker.com" which is a domain-global cookie. Users may need to flush out their old cookies by browsing: here, then here, then here, and finally here. (NOTE: You will need to log in again.)

If you really want to, you can also delete all forum cookies with this link.

The http://www.cellartracker.com/phone/ subsite now has cookies that will last for a year instead of just the current session. There is also a link to log out.

BEHIND THE SCENES
To be announced shortly in more depth, I now have my first official help in editorially reviewing the wine database. To better accomodate this, I cleaned up some administrative tools so that the site now has the concept of both an administrators and editors which I can easily add and remove from one central location.

< Message edited by Eric -- 4/6/2005 1:00:38 AM >


_____________________________

Cheers!
-Eric LeVine
Post #: 1
RE: v 3.0.1 release notes - 4/5/2005 10:29:12 PM   
PaulH

 

Posts: 251
Joined: 11/3/2003
From: San Francisco, CA USA
Status: offline

(in reply to Eric)
Post #: 2
RE: v 3.0.1 release notes - 4/5/2005 11:21:01 PM   
Eric

 

Posts: 5743
Joined: 10/10/2003
From: Seattle, WA
Status: offline
Paul, if you liked the original you will REALLY like the fact that I just added your '100 point penalty' for wines that are past their drinking window.

_____________________________

Cheers!
-Eric LeVine

(in reply to PaulH)
Post #: 3
RE: v 3.0.1 release notes - 4/6/2005 7:06:10 AM   
PaulH

 

Posts: 251
Joined: 11/3/2003
From: San Francisco, CA USA
Status: offline
quote:

ORIGINAL: Eric

Paul, if you liked the original you will REALLY like the fact that I just added your '100 point penalty' for wines that are past their drinking window.
Thanks, but I find that I really didn't miss it. It should be interesting to see what the rest of the user community thinks about it.

On changing the calcuation to regard 375s as 1/2 bottle (and 1.5l at 2 bottles) I'm not yet sure what I think. Take the case of a wine where you have 4 375s and another wine where you have 4 750s. Assume they have the same drinking window. 25% of the way through the window the index will be 0.5 for the first wine and 1.0 for the second one -- in essence saying that it twice as important to drink the 750 as the 375, when in fact, they are an equal amnout into the window and should be equal in ranking.

(in reply to Eric)
Post #: 4
RE: v 3.0.1 release notes - 4/6/2005 8:01:16 AM   
Eric

 

Posts: 5743
Joined: 10/10/2003
From: Seattle, WA
Status: offline
Paul, the bottle size stuff takes some getting used to, but for anyone that has 375ml, 750ml and 1.5L of the same wine the change is actually pretty dramatic and quite important.

If the 100 point penalty is controversial I can make it a toggle option.

_____________________________

Cheers!
-Eric LeVine

(in reply to PaulH)
Post #: 5
RE: v 3.0.1 release notes - 4/6/2005 9:03:20 PM   
1800kidney

 

Posts: 246
Joined: 7/17/2004
From: Birmingham, AL
Status: offline
The 100 point penalty is a sledge hammer...very effective, and the simplicity is probably not going to be improved...

I do have an algorithmic suggestion, based on a post on Jan 6, which would be more elegant and could let users have another degree of freedom. Haven't really thought it through for the new bell shaped, reverse Boltzian, and triple whammy aging curves, but for the simple linear case...

"I have added an if/then statement to my off-line calculation of AI that arbitrarily doubles their availability index:
If Today > Last Date, then AI = 2 * (Should have consumed - Consumed) "

This is a weighting function; simple default model is WT = 1. Since you now have a default window on the Drinkability report for Drinkability formula, how about adding another pull down window for Weighting factor? With WT default (=1), 2, 4, 8, 16, the general equation would become:

If Today > Last date, then AI = WT*(should have consumed - consumed)

Interested in seeing what you and Paul think of this; for now the 100 point slam is very functional, and not very subtle

David

(in reply to Eric)
Post #: 6
RE: v 3.0.1 release notes - 4/7/2005 12:11:29 AM   
ianlumb

 

Posts: 209
Joined: 3/4/2004
From: London, UK
Status: offline
I'd actually prefer the drinkability index to be discrete and only apply to physical bottles - so it would never go over the physical count of bottles you have left. Similarly for negatives, which don't really exist in a physical world.

I am not saying that their end result isn't desirable - having top priority wines at the top of the report and low priority wines at the bottom is great - but all of this could be accomplished by some simple sorting behind the scenes (days to drink from & days past drink by). After all, it's not as if the drinkability index was ever wrong - it was more a case of differentiating between similar drinkability indices where one was a higher priority and that wasn't being reflected mathematically. Some not-so-subtle use of colour/hot-cold flagging could accomplish the same thing.

Maybe i think this way because i'd really like a future drinkability curve for each wine at some point, and arbitrary fixes to single time points break all that and are 'unclean'.

Just my tuppence-worth.

(in reply to 1800kidney)
Post #: 7
RE: v 3.0.1 release notes - 4/7/2005 12:16:08 AM   
ianlumb

 

Posts: 209
Joined: 3/4/2004
From: London, UK
Status: offline
quote:

ORIGINAL: 1800kidney
"I have added an if/then statement to my off-line calculation of AI that arbitrarily doubles their availability index:
If Today > Last Date, then AI = 2 * (Should have consumed - Consumed) "


Indicdentally, i'd have the same objection to this - it's massaging something that should reflect real, physical bottles to reflect your personal sorting preferences. I would however have no objections to something like this, or the 100-sledgehammer, being applied behind the scenes at the db-level to accomplish the sorting goals.

(in reply to 1800kidney)
Post #: 8
RE: v 3.0.1 release notes - 4/15/2005 10:56:29 PM   
Eric

 

Posts: 5743
Joined: 10/10/2003
From: Seattle, WA
Status: offline
quote:

ORIGINAL: ianlumb
I'd actually prefer the drinkability index to be discrete and only apply to physical bottles - so it would never go over the physical count of bottles you have left. Similarly for negatives, which don't really exist in a physical world.
[...]
Maybe i think this way because i'd really like a future drinkability curve for each wine at some point, and arbitrary fixes to single time points break all that and are 'unclean'.

Ian, I just introduced a new option to turn off the penalty (look for radio buttons at the left). It is also turned off for the Excel web query.

_____________________________

Cheers!
-Eric LeVine

(in reply to ianlumb)
Post #: 9
RE: v 3.0.1 release notes - 4/18/2005 5:05:22 AM   
ianlumb

 

Posts: 209
Joined: 3/4/2004
From: London, UK
Status: offline
That's great. Works like a dream.

(in reply to Eric)
Post #: 10
Page:   [1]
All Forums >> [Cellar Talk] >> Release Notes >> v 3.0.1 release notes Page: [1]
Jump to:





New Messages No New Messages
Hot Topic w/ New Messages Hot Topic w/o New Messages
Locked w/ New Messages Locked w/o New Messages
 Post New Thread
 Reply to Message
 Post New Poll
 Submit Vote
 Delete My Own Post
 Delete My Own Thread
 Rate Posts




Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI

0.109