PuffinPlot issue tracker

Bug: a88/286e

ID : 286e60a2-fe2d-4b8d-90a4-e60abdc56021
Short name : a88/286e
Status : open
Severity : minor
Assigned :
Reporter : Pontus Lurcock <pont@talvi.net>
Creator : Pontus Lurcock <pont@talvi.net>
Created : Wed, 07 Nov 2012 20:09:31 +0000
Target : 1.5
Summary : Multiple fits per sample [247]

Comment: --------- Comment ---------
ID: a1a02451-a027-4c1d-8f12-68eb2dffff13
Short name: a88/286e/a1a
From: Pontus Lurcock <pont@talvi.net>
Date: Wed, 07 Nov 2012 20:09:31 +0000

JMG writes:
It would be a great advantage to be able to save and display
multiple line and GC fits performed on a single sample (and to
specify which one is included in site mean). Although it does
seem to be able to associate both a line and GC fit on one sample
it only records one with the site parameter file.

Comment: --------- Comment ---------
ID: f27870b1-f4d0-4058-b11e-8e38ce45c271
Short name: a88/286e/f27
From: Pontus Lurcock <pont@talvi.net>
Date: Fri, 27 Feb 2015 14:39:27 +0000

A further request along these lines from Russ Burmester, 2014-12-14:

RB>> And probably difficult with the existing file structure, but it
RB>> would be handy to be able to save both anchored and free line fits
RB>> without having to use two separate files.

RB>> The aim of being able to save both free and anchored fits is so the
RB>> fits with the same points selected and means derived from them can
RB>> be compared. This is no problem if one is satisfied with the points
RB>> included in the fits because the file can be cloned. However, one
RB>> of the advantages of your program is that it is interactive and
RB>> begs to be used to play with and learn from different fits. Instead
RB>> of a new format to accommodate this, would it be possible to link
RB>> two data files to synchronize them?

PL> Interesting ideas -- the notion of linked data files got me
PL> thinking, and on reflection it might not be so hard to support
PL> multiple synchronized suites within the same file, with some quick
PL> mechanism for switching between them. Implementation would still be
PL> a big job but perhaps not as complex as supporting multiple fits at
PL> the sample level.