Plus-Minus-Debate - Doerner Institut

1
The Plus-Minus-Debate
13.10.2013 15:43
Jonathan Ashley-Smith (Cambridge) to Andreas Burmester (Doerner Institut Munich)
[…] The statement “fluctuations of no more than ±10% RH per 24 hrs within
this range” is perfectly clear. The argument on page 3 describing
‘mathematical sense’ is as childish as it is misleading. The requirement is
that, within the range 40-60% , the rate of change shall not exceed 10% in
24hrs. This change may be a drop from 60% to 50% or a rise from 45% to 55%.
The humidity is not permitted go out of the range 50 ±10%RH, nor is it allowed
to change by “20% during one day”. I agree that the science behind the
specified rate of change is a bit obscure. But I do not believe that the specified
rate is damaging within the specified range.
You make reference to a specification of 50% RH and 20oC, saying that it
originates with ICOM. I cannot find such a specification. The ICOM website has
a section for current standards and guidelines:
http://icom.museum/professional-standards/standards-guidelines/
The most up-to-date reference to the environment comes in the ICOM
Guidelines for Loans (1974) which state that: “Relative humidity in normal
circumstances should be maintained at 54±4 per cent”. No set point for
temperature is mentioned. However it is suggested that: “The lamps used do
not raise the surface temperature of the loan more than 3°C above room
temperature.” This could be interpreted that ICOM-defined ‘stability’ allows a
temperature range of 3 degrees around an arbitrary set point. Your suggestion
(paragraph 10) that ranges smaller than ±5% RH should not be demanded is
already on the slippery slope of ‘relaxation’ from ICOM’s ±4 per cent! In
“Running a Museum: a practical handbook” (2004) ICOM now says:
“Risk management replaces rigid standards for the museum environment”
No one can argue with the statement “stable is safe”. However, one difficulty
is that no one can achieve absolute stability, that is to maintain a flat line
without any variation day-by-day or season-by-season. The bigger difficulty is
that many cultural heritage objects are housed in historic buildings where
approaching anything better than ±10% is out of the question. In such
situations the use of microclimate cases, or the removal of the objects to ‘safe’
gallery conditions, are not reasonable options. One is left with the question
“what is stable?”. Since the Pinakothek der Moderne [in Munich] achieves
±2.5% with sustainable ease, should this be the definition of stable? In which
case your minimum demand of ±5% is even further down the slippery slope.
So I think a relevant question is how big is the range that can still be called
‘stable’? One possible answer comes from the European Standard EN 15757
Jonathan Ashley-Smith and Andreas Burmester: The Plus-Minus-Debate (2013),
http://www.doernerinstitut.de/downloads/Plus-Minus-Debate.pdf
2
(devised and published with the help of conservators) which allows for ±10%.
[…]
06.11.2013 11:57
Dear Jonathan,
[…] your arguments against our Munich Positions, where I have been one of
the authors, are partly true. In regard to the Bizot Group I share most of their
views, however, I would not call it green. I think you have the same view on
this. My big concern, however, are the Bizot [Interim] Guidelines, where I
disagree in some formulations. I am totally aware, that referring to ICOM is in
some way stupid, however, this is done on the museum managing level
constantly and understood as a world-wide agreement on a specific value for
loans (20°C, 50%). It is written in the Munich Positions that this specific value
is irrelevant as long as the climate is stable. This stability was one of the early
demands of the Bizot specifications and has a long tradition in the loan
business. The later addition in the Bizot Guidelines as:
“a range of 40–60% and a stable temperature in the range 16–25°C with
fluctuations of no more than ±10% RH per 24 hours within this range”
is, as I think, misleading, as it does not tell you how many cycles of +-10% RH
are allowed within these 24 hours. To my experience, unlimited cycles of +-10
have nothing to do with stability in a conservator’s sense and is too much for
fragile and sensitive works of art on paper, for photographs, weakened
ethnographic material etc. It even might transform nowadays stable objects
into the fragile sorrows of tomorrow. It is obvious, that the formulation
perfectly allows to set up exhibitions in inappropriate climates. As you
experienced, inappropriate from a conservator’s view will rarely change the
overall planning of the head of an institution. In consequence, more battle and
more defence will be required in the future. This is the point of my main
concern. […]
With warm regards
Andreas
Jonathan Ashley-Smith and Andreas Burmester: The Plus-Minus-Debate (2013),
http://www.doernerinstitut.de/downloads/Plus-Minus-Debate.pdf
3
07.11.2013 14:18
Dear Jonathan,
I may come back to your comments regarding our ‘childish’ mathematical
statement. If you as Dario [Camuffo] based on work of our Polish friends
[Jakiela, Bratasz and Kozlowski]
speak about a Delta RH of 10, your are safe in the range of 30 to 70 %RH. Here,
the change is 10% instantly or within 24 hrs. If you, however, as the Bizots do,
speak about
“fluctuations of no more than +-10% RH per 24 hrs within this range“
is this at least misleading. It means: If you start at 50% - to our understanding
- you are moving in a target range of 40 to 60%. This view is supported by
Bratasz paper on page 016 in our conference volume where he quotes
EN15757:2010 as “Short-term fluctuation +-10 %RH […] The lower and upper
limits of the target range of RH fluctuations are determined as the 7th and 93rd
percentiles of the fluctuations recorded in the monitoring period respectively”.
From a mathematical point of view, this is perfectly clear.
So, the Bizot guidelines clearly define a target range of +-10% within a
specified range of 40 to 60% RH.
If I follow your interpretation, you have to argue with the absolute value:
“fluctuations of no more than 10% RH per 24 hrs within this range“.
Jonathan Ashley-Smith and Andreas Burmester: The Plus-Minus-Debate (2013),
http://www.doernerinstitut.de/downloads/Plus-Minus-Debate.pdf
4
If you do so, there is full agreement.
With regards
Andreas
11.11.2013 15:25
Dear Andreas,
I'm sorry not to have replied sooner. I was away […]. I'm glad that there is
quite a bit of common ground in our attitudes, yet a continuing understanding
that we will not always agree on everything.
best wishes
Jonathan
11.11.2013 14:55
this answer cant be serious. It is not the question whether we agree or not, it
is rather the question where you or I am wrong. There is a clear difference
between your definition Delta RH = 10 and the Bizot Interim Guidelines of +- 10
RH (equals to a Delta RH of 20) which needs explanation. Either you are right
or you are fighting for the wrong Guidelines. Or maybe you are right, and this
needs explanation too.
With regards
Andreas
11.11.2013 16:39
Dear Andreas,
my interpretation of “fluctuations of no more than +-10% RH per 24 hrs within
this range“ is " no fluctuation of more than plus 10% in 24hrs or more than
minus 10% in 24hrs anywhere in this range" which you have succinctly stated
as “fluctuations of no more than 10% RH per 24 hrs within this range“. We
agree.
Jonathan Ashley-Smith and Andreas Burmester: The Plus-Minus-Debate (2013),
http://www.doernerinstitut.de/downloads/Plus-Minus-Debate.pdf
5
Where we must continue to disagree (amicably) is whether the Bizot definition
is ambiguous and possibly misleading. I believe that you have to go out of your
way to interpret what is written as Delta RH =20 within 24 hrs.
In the Munich position document you say: "Do not ask about the meaning of
40–60 % ±10 % RH in 24 hours! " which allows you to speculate that the
"relative humidity is allowed to be between 30 and 50 % or 50 to 70 %". This
is something the Bizot statement clearly does not say. It says "with
fluctuations of no more than ±10% RH per 24 hours within this range". The
maximum allowed variation should not exceed 10% in either direction within
24hrs. The maximum possible variation is 20% but this cannot occur in a
single 24hr period. In an extreme situation the the RH could vary by as much
as 20% in a 48hr period while staying within the allowable range.
Jonathan
11.11.2013 17:06
Dear Jonathan,
geht doch! This is [in Germany] a fashioned way to say: Although we both
supposed that we shall not agree, you now do what I wanted. Thank you for
this. There is full agreement on point 1. Regarding the second point ‘within
this range’ you are right. I shall change this tomorrow on the internet. I this
case I do what you want, last but not least because I try to be fair to find a way
out of this struggle.
Regarding the last point: Whether I or you have to go out my or your way is still
not decided. This is why I sent you some days ago the following argument: ‘If
you start at 50% - to our understanding - you are moving in a target range of
40 to 60%. This view is supported by Bratasz paper on page 016 in our
conference volume where he quotes EN15757:2010 as “Short-term fluctuation
+-10 %RH […] The lower and upper limits of the target range of RH
fluctuations are determined as the 7th and 93rd percentiles of the fluctuations
recorded in the monitoring period respectively”’. From a mathematical point of
view, this is perfectly clear: +- never means that you go either + into one
direction or – in the other one. There is not much I remember from my time at
university at the mathematics department, but here Bratasz and I have the
same interpretation. It means that 86% of the RH values are to be found in a
window to be seen as frequency distribution around the mean value. I think,
there not much space left for any other interpretation. What do you think?
Sorry to bother you, Andreas
Jonathan Ashley-Smith and Andreas Burmester: The Plus-Minus-Debate (2013),
http://www.doernerinstitut.de/downloads/Plus-Minus-Debate.pdf
6
12.11.2013 16:48
Dear Andreas,
If I understand you correctly, you say that the plus-minus sign can never
indicate a choice between one direction or another. This is where we definitely
disagree. The symbol can mean a choice between a negative number or a
positive number, or it can mean a range about some central point. So its use in
the Bizot document to mean a change in one direction or a change in another
direction, is quite legitimate. I argue this at greater length in the attached
document [here added as post scriptum].
The sentence in EN15757:2010 that includes " Short-term fluctuation +-10
%RH " means that the fluctuation can go in either direction. It would be bizarre
if it could only go in one direction. I agree "that 86% of the RH values are to be
found in a window to be seen as frequency distribution around the mean
value", but what you would observe on a thermohygrograph chart would be the
RH sometimes varying in a positive direction away from the mean, and on
others moving in a negative direction away from the mean.
Jonathan
post scriptum on ±
My first memory of the plus minus sign comes from its use with square
roots, where a number has both a positive and a negative root. √16 = ± 4
the square root of 16 can be either plus 4 or minus 4
I next came across it in describing errors in measurement. The length
of this object is 60 mm ± 0.5mm
It was much later that I came across its use to describe a range,
50%±5%, meaning within the range 45-55%.
So, if there is no number immediately in front of the ± sign, it indicates
that the number immediately following the ± sign can be viewed as a
positive number or as a negative number. This is the way that it is used
in the Bizot definition.
If there is a number immediately in front of the ± sign then the number
immediately following the ± sign indicates the size of an uncertainty or
the range in a specification.
In support of my memory the source of all wisdom Wikipedia says this
of the plus-minus sign:
Jonathan Ashley-Smith and Andreas Burmester: The Plus-Minus-Debate (2013),
http://www.doernerinstitut.de/downloads/Plus-Minus-Debate.pdf
7
“The plus-minus sign (±) is a mathematical symbol with multiple
meanings.
In mathematics, it generally indicates a choice of exactly two possible
values, one of which is the negation of the other.
In experimental sciences, the sign commonly indicates the confidence
interval or error in a measurement, often the standard deviation or
standard error.
The sign may also represent an inclusive range of values that a reading
might have.”
the German Wikipedia says much the same:
Das Plusminuszeichen (±) und das Minuspluszeichen (∓) sind Zeichen
aus der Mathematik. Mit ihnen kann zum einen ausgedrückt werden,
dass ein Term in einem mathematischen Ausdruck sowohl positiv als
auch negativ sein kann (so bedeutet „±5“, dass die betrachtete Zahl +5
oder −5 sein kann).
12.11.2013 17:05
Dear Jonathan,
there is total agreement on this. In other terms the +- describes a unbiased
window or range within the RH value can freely fluctuate. If we cut the window
by saying ‘40 – 60 +- 10% within this range’, however, we can not speak about
a normal frequency distribution, it is rather biased. How to do this practically
in air conditioning is another question and is not within our debate.
Clearly, the +- sign indicates that there a values above and below the mean. If
this would not be the case, the mean would move. If we want to have a stable
RH the mean must be stable in a long term.
As in every frequency distribution, and due to external influences, a value can
be on the negative side and hours later or even in the next moment on the
positive side. This means that for the +- 10% Bizot case the value can in a
worst case move within the 24 hours from (mean – 10 %) to (mean + 10%). So
the difference between these two values is 20. Do we agree on this?
If so, remember the Camuffo plot I sent you [on 07.11.2013]. If delta RH is 20%
we are in the case of an instantaneous change in the area of irreversible
Jonathan Ashley-Smith and Andreas Burmester: The Plus-Minus-Debate (2013),
http://www.doernerinstitut.de/downloads/Plus-Minus-Debate.pdf
8
change, in the case of a slow change lasting 24 hours for a mean of around 50
on the border of the (ir)reversible change. This is true for wood Camuffo is
referring to, it is might be not true for sensitive photographic paper. At that
point we clearly have to speak about increased risks, do we agree on this?
In consequence, to avoid biased situations and to reduce risk, we might agree
on ‘fluctuations of no more than 10% RH per 24 hrs within this range’. This is
equivalent to RH mean +- 5%? Or is it too early to make this step? There is
more to be said in a later email.
With regards
Andreas
12.11.2013 19:29
Dear Andreas,
before your promised next email, I have to say that I still do not share your
interpretation of the Bizot statement. Bear in mind that the statement is a
specification not an analysis of past conditions.
In the Bizot statement there is no mention of a mean, only of a range. This is
40-60. Staying within this range is the Bizot definition of stable. Conditions
could stay at 41% plus or minus 1% (nowhere near the midpoint of 50%) and
still be in this range. The humidity could be 60% one day and 40% at some
later stage. However it could not be 60% and then 40% within 24hours.
This is because a maximum rate of change is specified. This maximum
fluctuation rate is either a rise of 10% in 24hrs or a fall of 10% in 24 hrs (either
plus or minus 10%). The rise could be 40%-50% or 45%-55% the fall could be
59% to 49% or 53% to 43%. There is no mention of a mean or of a specific set
point. There is only a range, defined by an upper and lower boundary, and a
maximum rate of change within this range.
Looking at the diagrams: If the conditions stay within this range and the rate of
change in either direction does not exceed 10% in 24 hrs the induced strain
stays in the reversible response range.
I agree that if the Bizot specification did allow a change of 20% in 24 hrs the
risk of damage would be much higher. However it does not.
Jonathan Ashley-Smith and Andreas Burmester: The Plus-Minus-Debate (2013),
http://www.doernerinstitut.de/downloads/Plus-Minus-Debate.pdf
9
Incidentally I think that the diagrams that you refer to as from Camuffo
originate from Jakiela, Bratasz and Kozlowski 2008, Wood Science and
Technology.
Jonathan
13.11.2013 9:54
Dear Jonathan,
this evening I shall be on my way to England. No fear, I shall not knock on you
door and I have no computer with me. So, no further email about our topic. If I
try to summarize, there is full agreement on all details around the Bizot
Interim Guidelines. Including that Camuffos diagrams are from the Polish
group which I point out in my PPPs but did not in my email.
However, the way how we read the [Bizot Interim] Guideline is different. […]
there must be something misleading with the Guideline. We have spotted the
+- where the main problem is. And, as you said, my formulation
“fluctuations of no more than 10% RH per 24 hrs within the range of 40 to
60%“
is much clearer and would keep the risk manageable. If we agree on this, the
second problem we have to spot is the ‘range versus mean’ topic. Both views
allow a change from 45 to 55%, we would stay in the range as well as around
the mean of 50%. Whether a corridor can be managed by air conditioning
systems is unclear for all the existing machinery, it might be possible for
future equipment. Whether we have to expect an increase or decrease in cost,
is under discussion, if you use the machines different from their original
design. This discussion has to be left for the future.
As I understood on a meeting of the Leipziger Kreis [the Directors Conference
of Leading German Museums] on Monday, the conservation world is now
asked by the Bizots to sort out the muddle. I think, my proposal could be a way
out of at least an ambiguous situation. Whether the conservation and museum
world finally will accept the improved formulation, is another topic. I would not
for sensitive objects, you might, but you are in the lucky situation not to carry
the risk for loans anymore as you had to do in the past. Lucky person!
Would you agree that I put our +- discussion on our website? […] I do think,
that this would help in the discussion.
With warm regards Andreas
Jonathan Ashley-Smith and Andreas Burmester: The Plus-Minus-Debate (2013),
http://www.doernerinstitut.de/downloads/Plus-Minus-Debate.pdf