A forum for people with knowledge of the Bible in its original languages to discuss its manuscripts and textual history from the perspective of historic evangelical theology.
Klaus Wachtel from the INTF in Münster reports that version 1.0 of the on-line software 'Genealogical Queries' including a guide in English and German is available here.
The problem I have with this method is that I still don't understand it. I have read the articles, some twice and trice, but it doesn't help. I understand pre-genealogical evidence and I understand local stemmata, but I don't understand how the program works. It is a black box. Now I read in Wachtel's article about new parameters like "connectivity". Perhaps, as a suggestion, it might be a good idea to make up say ten (artificial) variants with ten (artificial) witnesses and show exactly how one gets to the global stemma and what happens in-between. I still have the suspicion that a lot of prior knowledge is involved and also that circularity is a problem.
Perhaps the method is sound, but I am unable to evaluate it.
Well, I have a doctoral dissertation coming in which I take an issue with the whole method. It is based on problematic (IMO) assumptions on the early scribal habits. The method seems to work well with later witnesses (Byz) but not with the witnesses of the earliest period. The whole "black box" is really just a mathematical probability statistics that gives the impression of being sound science when in reality it is a subjective evaluation of a given variant location on the basis of subjective evaluations of other locations (the iterative process). This branching heuristic/fuzzy logic system taken from biology and computer science does not work well with human beings copying a text :) I remain spectical about the value of the whole thing.
"The method seems to work well with later witnesses (Byz) but not with the witnesses of the earliest period."
I think this is an important statement. My gut feeling tells me that this is probably right. (It then is clear that the method works comparatively well with the Catholics.)
Well, to answer Wieland's question, before shifting to theology, I used to be a computer engineer with knowledge of fuzzy logic systems :)
Seriously, if you read everything Mink has written about the method, you should be able to pick it up too.
Actually, it is an article in my dissertation that deals with early scribal habits. Mink has given his views on this topic quite eloquently in his long article in Studies in Stemmatology II.
Timo Flink: "It [= Coherence based genealogical method] is based on problematic (IMO) assumptions on the early scribal habits. [1] The method seems to work well with later witnesses (Byz) but not with the witnesses of the earliest period... [2]"This branching heuristic/fuzzy logic system taken from biology and computer science does not work well with human beings copying a text."
I'm really looking forward to examining the material case for early scribal habits vs. later scribal habits.
As it stands right now, you should have a hard time to defend [1] and [2] as the later witnesses (Byz) seem to have also been copied by human beings, at the least the ones that I've been looking at.
the later period of transmission was more controlled than the early one (I think we can agree on that?). Yes, later scribes did mistakes, but the coherence overall is far better attested in later witnesses than in the very early ones. My [2] is a statement about the earliest period. It is simply so (IMO) that Mink's assumptions and the early scribal habits do not seem to fit the same glove whereas the later scribal habits seem to do so. I'm sure Mink will disagree as it is crucial to his theory that early and later scribal habits are essentially similar. I take a small consolation on the fact that J.K. Elliott, after reading the preliminary version, agreed with me :)
It should be noted that I worked with the method as it was by the time the ECM in Catholics came out. I'm sure future developments in CBGM may require a re-evaluation of it.
Timo Flink: "My [2, i.e. '...does not work well with human beings copying a text'] is a statement about the earliest period."
So your statement is not about human beings, but about human beings (copying a text) from the earliest period. So when does the earliest period end?
IMO you need two establish two parameters when it comes to make a case for early copying as essentially different from later copying practises. (Well, the practises in terms of material developments and social settings certainly changed. But what effect did these changes have on the overall result of the copying effort?). Anyway, these are the parameters:
(1) A good model on the ratio of preserved and lost copies during the course of time. And a related good guess on the effect that changing ratios may have on the overall picture. (2) A good sample from both periods where you draw the line, resulting in a body of empirical data that allow to compare different periods. This, I'm afraid, involves transcribing/collating a couple of dozend "unknown" minuscules in substantial portions of text.
This would be a very welcome study, clearing up some (to me largely) unknown territory. And I'm seriously open to compelling evidence. In the meantime I think Mink's general reasoning stands: The overall coherence in the later part of the tradition is much stronger, because many more close relatives from the later periods (as opposed to the ealier periods) have survived. This is the challenge.
"Now I read in Wachtel's article about new parameters like "connectivity"." For connective variants cf. Studies in Stemmatology II 29 and 54/55.
"I still have the suspicion that a lot of prior knowledge is involved and also that circularity is a problem." For circularity cf. Studies in Stemmatology II 25. Circularity must be controlled, prior knowledge may be ascertained or falsified, or uncertainty may remain.
Timo Flink:
"It is based on problematic (IMO) assumptions on the early scribal habits." "It is simply so (IMO) that Mink's assumptions and the early scribal habits do not seem to fit the same glove whereas the later scribal habits seem to do so. I'm sure Mink will disagree as it is crucial to his theory that early and later scribal habits are essentially similar."
I did not make a special assumption on early scribal habits, and I do not know "the" early scibal habits but only habits of several scribes for the vast majority of early manuscripts is lost. The 4 assumptions (Studies in Stemmatology II 25) are only considered more probable than their contrary and they do not exclude anything what might happen. My statements are wrong if the contrary to exactly these statements is more probable. These assumptions are used only in order to have a model. Cases of transmission not according with this model will become evident. Trying the Gen. Queries facility out you can easily find examples.
"The method seems to work well with later witnesses (Byz) but not with the witnesses of the earliest period."
This is not accurate. There are clear differences between witnesses with earlier text (not depending on the date of the manuscript) and witnesses with later texts. Compare please the different results in the potential ancestors module and the different positions of those witnesses in the textual flow diagrams. Differences may be due to the large number of missing links or less controlled copying.
"The whole "black box" is really just a mathematical probability statistics that gives the impression of being sound science when in reality it is a subjective evaluation of a given variant location on the basis of subjective evaluations of other locations (the iterative process)."
Mathematical probability statistics is not involved in any procedure of the CBGM. Philological reasoning and using internal criteria has subjective elements to a certain degree. I do not know how we can avoid this totally. But utilising the CBGM we can get a better controll (overall view on what we are doing). Should we abandon this? The level of evaluation of readings is one of three levels of falsifiability (the others: level of facts and level of procedures and their rules, Studies in Stemmatology II 76).
"This branching heuristic/fuzzy logic system taken from biology and computer science does not work well with human beings copying a text"
The system is not taken from biology or computer science. Fuzzy logic is not used in any programme.
thanks to Ulrich and Gerd for input. I appreaciate this.
Perhaps I need to explain what I meant by "heuristic/fuzzy logic" in this instance. The iterative process works the same way as certain fuzzy logic systems. Hence my politically incorrect terminology here on the web :) I do not use these labels in my dissertation at all.
I agree that the subjectivity cannot be totally avoided. It's always there, somehow.
However, after thinking through the data we have of the early scribal habits (2nd-3rd c), I became convinced that the assumptions do not work well enough. It seems to me on the basis of the data that while the 4 assumptions allow room for the contrary in some places, in actual fact the contrary takes place too often. Obviously this may be partially because of so much has been lost, but that does not change in my opinion the problem of sorting out the early witnesses by CBGM. There are simply too many problems with the early scribes and the texts known from the early period. I have no idea how the system could be enhanced to work around this. At least not yet. My point in the dissertation is simply to check whether CBGM can reach the second century text despite problems with early scribal habits. (I can delete the "the" :)). So far I am not convinced it can. At the moment I have concluded that CBGM works with post-fourth century, but not before it.
As I see it, the iterative process does involve mathematical probabilities (though they are not seen as such). The prior/posterior sorting is a form of statistics. Using that info iteratively to reprocess the system so that the system tries to find an equilibrium is what neural nets do in some fuzzy logic systems on the basis of certain parameters, which can be quantified as mathematical probabilities. Pardon me for using these labels on this discussion. I won't use them again, as this appears to cause a misunderstanding.
Oh, I agree that using "Byz" is inaccurate. My apologies. I mean witnesses with later texts, which were copied with more control. I wonder if someone could come up a good sorthand for that...
I may add that no, we should not abandon the CBGM, but we should be aware of its limitations in relation to the very early period of transmission.
That's why I will propose that the system should be used to sort out all post-fourth century witnesses to find the genealogically closests witnesses to the hypothetical initial text, but leave the very early witnesses out because of problematic scribal habits. Then we simply do a reasoned eclecticism using witnesses that cannot be sorted well enough and witnesses with, say, first five generations removed from the initial text.
In fact, I will offer an experiment with that in my dissertation to see if that helps. We'll see.
I think that it is not necessarily different scribal habits, but that the genealogical relationship between the very early witnesses is possibly not strong enough. Just think of the codices 02, 03 and 05 in the Gospels as an example. The software will probably find a place in a stemma for them, but is this then real?
There are so many missing links, so many possibilities (including some kind of recension), that I am wondering if a possible stemma really represents reality or isn't just some kind of imagination.
Since I don't know enough about the catholics to say something definite, I am awaiting the application of the method to the Gospels. I am really looking forward to this.
yes, that is also a problem, but as I said, the early scribal habits are also a problem. Well, in so far as they are known from the available early witnesses. This btw relates mostly to the Gospels :)
Thanks for some clarification and less controversial prose, Timo.
Now I have, at least some idea about the time period where you want to draw a line between earliest and later scribal activities. However, there is still a lot more to address in order to make a satisfactory case in that regard. You may not want to do it here, but those questions linger.
You say: "However, after thinking through the data we have of the early scribal habits (2nd-3rd c), I became convinced that the assumptions do not work well enough." And: "At the moment I have concluded that CBGM works with post-fourth century, but not before it."
The elephant in the room is this: What about fourth century witnesses 01 and 03? Are they qualified for inclusion or not? Perhaps the work of scribe D of 01 might be included but not the work of scribe A (see Jonkind, Scribal Habits of Codex Sinaiticus, 2007). Unfortunately, scribe A contributed most of the NT.
Moreover, Jongkind even goes on to state: "Sinaiticus may be evidence for the non-existence of the 'fourth-century barrier'... One may wonder, therefore, whether it is correct to split the history of textual transmission into rigotously and non-rigorously controlled periods." (p. 246).
In short, there is a lot more work necessary to assemble data from other periods in order to maintain a distinction that may only be an artificial one tying in with some gut feeling but not supported by actual evidence.
Yep, I tend to be controversial at times :) [my way of getting debates going, sadly]
I think 03 needs to be included as its text goes to pre-fourth c. date (I do so in my diss.). As for 01, that is more like sitting in the fence, but probably it should be included as well. These two are of course interesting exceptions to the general tendency that post-fourth century witnesses are more controlled.
Are you perhaps thinking that if 01 and 03 need to be included and some later witnesses are close to 01 and 03, I don't have a case?
"What is the minimum number of witnesses and variants for the method to work? Roughly?"
Theoretically, the minimum number is one variant passage and two witnesses, but you would not need the method. It is the more useful the more passages and witnesses there are.
"There are so many missing links, so many possibilities (including some kind of recension), that I am wondering if a possible stemma really represents reality or isn't just some kind of imagination."
A stemma is a hypothesis. Hypothesis is not reality. We need a hypothesis where we cannnot find out "how it really was". The particular validity of any hypothesis must be defined. In this case, it is the simplest genealogical structure of the preserved (!) tradition based on our philological and critical reasoning. Interpreting this structure in terms of history is another job.
Timo Flink:
"However, after thinking through the data we have of the early scribal habits (2nd-3rd c), I became convinced that the assumptions do not work well enough. It seems to me on the basis of the data that while the 4 assumptions allow room for the contrary in some places, in actual fact the contrary takes place too often."
The 4 assumptions present the preferred model in the case of closely related witnesses. When cases do not accord with this model and, thus, there aro no close relatives, the 4 assumptions do not affect the results. Again, please check the potential ancestors module and the comparison modul in order to see the different results e.g. for P72 (special situation in 2Pt) and P100. It is not the method but the fragmentation that delimits genealogical significance. Also in later times there are manuscripts with text which do not conform to the 4 assumptions (e.g. 044, especially in 2Pt, or 631).
This is excellent. Now I can plot my family tree.
ReplyDeleteYes, you might be related to Codex Vaticanus.
ReplyDeleteThe problem I have with this method is that I still don't understand it. I have read the articles, some twice and trice, but it doesn't help. I understand pre-genealogical evidence and I understand local stemmata, but I don't understand how the program works. It is a black box. Now I read in Wachtel's article about new parameters like "connectivity".
ReplyDeletePerhaps, as a suggestion, it might be a good idea to make up say ten (artificial) variants with ten (artificial) witnesses and show exactly how one gets to the global stemma and what happens in-between.
I still have the suspicion that a lot of prior knowledge is involved and also that circularity is a problem.
Perhaps the method is sound, but I am unable to evaluate it.
Well, I have a doctoral dissertation coming in which I take an issue with the whole method. It is based on problematic (IMO) assumptions on the early scribal habits. The method seems to work well with later witnesses (Byz) but not with the witnesses of the earliest period. The whole "black box" is really just a mathematical probability statistics that gives the impression of being sound science when in reality it is a subjective evaluation of a given variant location on the basis of subjective evaluations of other locations (the iterative process). This branching heuristic/fuzzy logic system taken from biology and computer science does not work well with human beings copying a text :) I remain spectical about the value of the whole thing.
ReplyDelete"Well, I have a doctoral dissertation coming"
ReplyDeleteWow! I must read it!
"The method seems to work well with later witnesses (Byz) but not with the witnesses of the earliest period."
I think this is an important statement. My gut feeling tells me that this is probably right. (It then is clear that the method works comparatively well with the Catholics.)
But I am open to anything.
How did you learn how the method really works?
Well, to answer Wieland's question, before shifting to theology, I used to be a computer engineer with knowledge of fuzzy logic systems :)
ReplyDeleteSeriously, if you read everything Mink has written about the method, you should be able to pick it up too.
Actually, it is an article in my dissertation that deals with early scribal habits. Mink has given his views on this topic quite eloquently in his long article in Studies in Stemmatology II.
TF: "Well, I have a doctoral dissertation coming in which I take an issue with the whole method."
ReplyDeleteI look forward to reading your work too!
Timo Flink:
ReplyDelete"It [= Coherence based genealogical method] is based on problematic (IMO) assumptions on the early scribal habits.
[1] The method seems to work well with later witnesses (Byz) but not with the witnesses of the earliest period...
[2]"This branching heuristic/fuzzy logic system taken from biology and computer science does not work well with human beings copying a text."
I'm really looking forward to examining the material case for early scribal habits vs. later scribal habits.
As it stands right now, you should have a hard time to defend [1] and [2] as the later witnesses (Byz) seem to have also been copied by human beings, at the least the ones that I've been looking at.
Ulrich Schmid
to Ulrich,
ReplyDeletethe later period of transmission was more controlled than the early one (I think we can agree on that?). Yes, later scribes did mistakes, but the coherence overall is far better attested in later witnesses than in the very early ones. My [2] is a statement about the earliest period. It is simply so (IMO) that Mink's assumptions and the early scribal habits do not seem to fit the same glove whereas the later scribal habits seem to do so. I'm sure Mink will disagree as it is crucial to his theory that early and later scribal habits are essentially similar. I take a small consolation on the fact that J.K. Elliott, after reading the preliminary version, agreed with me :)
It should be noted that I worked with the method as it was by the time the ECM in Catholics came out. I'm sure future developments in CBGM may require a re-evaluation of it.
well, it's not finished yet, so...
Timo Flink:
ReplyDelete"My [2, i.e. '...does not work well with human beings copying a text'] is a statement about the earliest period."
So your statement is not about human beings, but about human beings (copying a text) from the earliest period. So when does the earliest period end?
IMO you need two establish two parameters when it comes to make a case for early copying as essentially different from later copying practises. (Well, the practises in terms of material developments and social settings certainly changed. But what effect did these changes have on the overall result of the copying effort?). Anyway, these are the parameters:
(1) A good model on the ratio of preserved and lost copies during the course of time. And a related good guess on the effect that changing ratios may have on the overall picture.
(2) A good sample from both periods where you draw the line, resulting in a body of empirical data that allow to compare different periods. This, I'm afraid, involves transcribing/collating a couple of dozend "unknown" minuscules in substantial portions of text.
This would be a very welcome study, clearing up some (to me largely) unknown territory. And I'm seriously open to compelling evidence. In the meantime I think Mink's general reasoning stands: The overall coherence in the later part of the tradition is much stronger, because many more close relatives from the later periods (as opposed to the ealier periods) have survived. This is the challenge.
Ulrich Schmid
Wieland Willker:
ReplyDelete"Now I read in Wachtel's article about new parameters like "connectivity"."
For connective variants cf. Studies in Stemmatology II 29 and 54/55.
"I still have the suspicion that a lot of prior knowledge is involved and also that circularity is a problem."
For circularity cf. Studies in Stemmatology II 25. Circularity must be controlled, prior knowledge may be ascertained or falsified, or uncertainty may remain.
Timo Flink:
"It is based on problematic (IMO) assumptions on the early scribal habits."
"It is simply so (IMO) that Mink's assumptions and the early scribal habits do not seem to fit the same glove whereas the later scribal habits seem to do so. I'm sure Mink will disagree as it is crucial to his theory that early and later scribal habits are essentially similar."
I did not make a special assumption on early scribal habits, and I do not know "the" early scibal habits but only habits of several scribes for the vast majority of early manuscripts is lost.
The 4 assumptions (Studies in Stemmatology II 25) are only considered more probable than their contrary and they do not exclude anything what might happen. My statements are wrong if the contrary to exactly these statements is more probable.
These assumptions are used only in order to have a model. Cases of transmission not according with this model will become evident. Trying the Gen. Queries facility out you can easily find examples.
"The method seems to work well with later witnesses (Byz) but not with the witnesses of the earliest period."
This is not accurate. There are clear differences between witnesses with earlier text (not depending on the date of the manuscript) and witnesses with later texts. Compare please the different results in the potential ancestors module and the different positions of those witnesses in the textual flow diagrams. Differences may be due to the large number of missing links or less controlled copying.
"The whole "black box" is really just a mathematical probability statistics that gives the impression of being sound science when in reality it is a subjective evaluation of a given variant location on the basis of subjective evaluations of other locations (the iterative process)."
Mathematical probability statistics is not involved in any procedure of the CBGM. Philological reasoning and using internal criteria has subjective elements to a certain degree. I do not know how we can avoid this totally. But utilising the CBGM we can get a better controll (overall view on what we are doing). Should we abandon this? The level of evaluation of readings is one of three levels of falsifiability (the others: level of facts and level of procedures and their rules, Studies in Stemmatology II 76).
"This branching heuristic/fuzzy logic system taken from biology and computer science does not work well with human beings copying a text"
The system is not taken from biology or computer science. Fuzzy logic is not used in any programme.
Gerd Mink
Gerd, given a typical NT situation: What is the minimum number of witnesses and variants for the method to work? Roughly?
ReplyDeletePS: Ok, "connectivity" isn't new. I just forgot it. I promise to re-read everything once again.
thanks to Ulrich and Gerd for input. I appreaciate this.
ReplyDeletePerhaps I need to explain what I meant by "heuristic/fuzzy logic" in this instance. The iterative process works the same way as certain fuzzy logic systems. Hence my politically incorrect terminology here on the web :) I do not use these labels in my dissertation at all.
I agree that the subjectivity cannot be totally avoided. It's always there, somehow.
However, after thinking through the data we have of the early scribal habits (2nd-3rd c), I became convinced that the assumptions do not work well enough. It seems to me on the basis of the data that while the 4 assumptions allow room for the contrary in some places, in actual fact the contrary takes place too often. Obviously this may be partially because of so much has been lost, but that does not change in my opinion the problem of sorting out the early witnesses by CBGM. There are simply too many problems with the early scribes and the texts known from the early period. I have no idea how the system could be enhanced to work around this. At least not yet. My point in the dissertation is simply to check whether CBGM can reach the second century text despite problems with early scribal habits. (I can delete the "the" :)). So far I am not convinced it can. At the moment I have concluded that CBGM works with post-fourth century, but not before it.
As I see it, the iterative process does involve mathematical probabilities (though they are not seen as such). The prior/posterior sorting is a form of statistics. Using that info iteratively to reprocess the system so that the system tries to find an equilibrium
is what neural nets do in some fuzzy logic systems on the basis of certain parameters, which can be quantified as mathematical probabilities. Pardon me for using these labels on this discussion. I won't use them again, as this appears to cause a misunderstanding.
Oh, I agree that using "Byz" is inaccurate. My apologies. I mean witnesses with later texts, which were copied with more control. I wonder if someone could come up a good sorthand for that...
anyways, I am thankful for your comments.
I may add that no, we should not abandon the CBGM, but we should be aware of its limitations in relation to the very early period of transmission.
ReplyDeleteThat's why I will propose that the system should be used to sort out all post-fourth century witnesses to find the genealogically closests witnesses to the hypothetical initial text, but leave the very early witnesses out because of problematic scribal habits. Then we simply do a reasoned eclecticism using witnesses that cannot be sorted well enough and witnesses with, say, first five generations removed from the initial text.
In fact, I will offer an experiment with that in my dissertation to see if that helps. We'll see.
I think that it is not necessarily different scribal habits, but that the genealogical relationship between the very early witnesses is possibly not strong enough. Just think of the codices 02, 03 and 05 in the Gospels as an example. The software will probably find a place in a stemma for them, but is this then real?
ReplyDeleteThere are so many missing links, so many possibilities (including some kind of recension), that I am wondering if a possible stemma really represents reality or isn't just some kind of imagination.
Since I don't know enough about the catholics to say something definite, I am awaiting the application of the method to the Gospels. I am really looking forward to this.
to Wieland,
ReplyDeleteyes, that is also a problem, but as I said, the early scribal habits are also a problem. Well, in so far as they are known from the available early witnesses. This btw relates mostly to the Gospels :)
Thanks for some clarification and less controversial prose, Timo.
ReplyDeleteNow I have, at least some idea about the time period where you want to draw a line between earliest and later scribal activities. However, there is still a lot more to address in order to make a satisfactory case in that regard. You may not want to do it here, but those questions linger.
You say: "However, after thinking through the data we have of the early scribal habits (2nd-3rd c), I became convinced that the assumptions do not work well enough."
And: "At the moment I have concluded that CBGM works with post-fourth century, but not before it."
The elephant in the room is this: What about fourth century witnesses 01 and 03? Are they qualified for inclusion or not? Perhaps the work of scribe D of 01 might be included but not the work of scribe A (see Jonkind, Scribal Habits of Codex Sinaiticus, 2007). Unfortunately, scribe A contributed most of the NT.
Moreover, Jongkind even goes on to state: "Sinaiticus may be evidence for the non-existence of the 'fourth-century barrier'... One may wonder, therefore, whether it is correct to split the history of textual transmission into rigotously and non-rigorously controlled periods." (p. 246).
In short, there is a lot more work necessary to assemble data from other periods in order to maintain a distinction that may only be an artificial one tying in with some gut feeling but not supported by actual evidence.
Ulrich Schmid
Yep, I tend to be controversial at times :) [my way of getting debates going, sadly]
ReplyDeleteI think 03 needs to be included as its text goes to pre-fourth c. date (I do so in my diss.). As for 01, that is more like sitting in the fence, but probably it should be included as well. These two are of course interesting exceptions to the general tendency that post-fourth century witnesses are more controlled.
Are you perhaps thinking that if 01 and 03 need to be included and some later witnesses are close to 01 and 03, I don't have a case?
Good questions though. Thx.
Wieland Willker:
ReplyDelete"What is the minimum number of witnesses and variants for the method to work? Roughly?"
Theoretically, the minimum number is one variant passage and two witnesses, but you would not need the method. It is the more useful the more passages and witnesses there are.
"There are so many missing links, so many possibilities (including some kind of recension), that I am wondering if a possible stemma really represents reality or isn't just some kind of imagination."
A stemma is a hypothesis. Hypothesis is not reality. We need a hypothesis where we cannnot find out "how it really was". The particular validity of any hypothesis must be defined. In this case, it is the simplest genealogical structure of the preserved (!) tradition based on our philological and critical reasoning. Interpreting this structure in terms of history is another job.
Timo Flink:
"However, after thinking through the data we have of the early scribal habits (2nd-3rd c), I became convinced that the assumptions do not work well enough. It seems to me on the basis of the data that while the 4 assumptions allow room for the contrary in some places, in actual fact the contrary takes place too often."
The 4 assumptions present the preferred model in the case of closely related witnesses. When cases do not accord with this model and, thus, there aro no close relatives, the 4 assumptions do not affect the results. Again, please check the potential ancestors module and the comparison modul in order to see the different results e.g. for P72 (special situation in 2Pt) and P100. It is not the method but the fragmentation that delimits genealogical significance. Also in later times there are manuscripts with text which do not conform to the 4 assumptions (e.g. 044, especially in 2Pt, or 631).
Gerd Mink
Thanks Gerd, I did not know that about 631.
ReplyDelete