Author Topic: Houston - We Have a Problem  (Read 8817 times)

DerekSmith

  • IGKT Member
  • Sr. Member
  • *****
  • Posts: 1521
  • Knot Botherer
    • ALbion Alliance
Re: Houston - We Have a Problem
« Reply #15 on: December 11, 2006, 10:46:24 PM »
Derek,

While this is the IGKT, we are not necessarily the foremost experts in knotting.

Nothing to stop us pushing our field forward though is there?

Quote
The frustration in this matter has caused some emotional outbursts at times.  I admit that I often will not read such descriptions.  The probability of success is too low, and the probability of error is too high.

I totally agree when we use rich unstructured English.  Even Dan's very clear instructions relating to the 'red end' were ambiguous - "pass the red end over itself then back under itself".  This can be interpreted in two ways 1. passed right then back under itself or 2. passed left then forward, back under itself.  The second way forms the overhand knot which is the only knot that can be made in-line, so this is probably the form Dan meant.  Using unconstrained English leaves the probability of error way too high.

Quote
The engineering fields do have a long-established method of communicating three-dimensional information.  It is drawing and diagrams.  I see no reason it need be different with knots.

And these used to be drawn using pen and paper and sent through the post.  Today they are drawn with AutoCAD and sent and displayed through the internet.  But guess what, behind those drawings is a text based language which describes exactly every line, curve, shade etc..  Those drawings are communicated in text files with an eps structure - encapsulated post script (or propriatory equivalent).  Strangely enough, restricting the number of words we can use and the way we can use them actually eliminates ambiguity, that why a CAD drawing will always display the same, there is no room for ambiguity in the text that describes how to recreate the drawing.

Quote
Just about any computer has at least a rudimentary draw/paint program.  I'd wager that a higher percentage of persons could successfully produce a graphical representation of knots than could describe them with text.

Most of us can draw with a pen and paper, but have you ever tried to draw with "Paint" ?  No matter how careful you are the result looks like a 2 year old did it.  Consequently we all try to get by with words rather than expose the world to our Painter works of 'art' and ourselves to rigorous jesting.

We will always need pictures, but I believe we also need a simple and unambiguous text based system that we can fall back on when needed, we just need to agree on one.

Derek

PatDucey

  • Sr. Member
  • *****
  • Posts: 287
  • IGKT Pacific Americas Branch
Re: Houston - We Have a Problem
« Reply #16 on: December 12, 2006, 02:08:36 AM »
Derek,

I am an AutoCAD drafter, and the templates I design to draw complex Turks Heads use about the same computer memory as a JPEG.  Some of the drawings I work with in the shipyard can get HUGE, but the simple stuff for knots are quite reasonable.  Typically I insert the AutoCAD into a Word document so it can be opened by almost anyone.  Whether it is a .doc, a .jpg, or a .dwg, it will take about the same computer memory to store an individual knot.

Before I was a drafter, I submitted to Knotting Matters a Paint drawing, which now that I look back at it, does look a bit rough.  But it was published, and it clearly conveyed the idea I was trying to get across.  AutoCAD cost about $1000 (US) a license, so it is also quite cost prohibitive.  There are many marvelous drafting tools that span from Paint to AutoCAD, and this is an area where you get what you pay for.  If you need the whiz-bang tools, AutoCAD is it.  For most knots, there are many fine alternatives.  I love the knots that Willeke draws, they are clear and concise, and I don't think that's AutoCAD she's using. 

Also, the text instructions behind the AutoCAD drawing are stupid.  In effect it says "draw a black line with width 2 from pt A to pt B".  Each line is a distinct entity, with no continuity if the line is broken.  Each separate entity is also drawn in the order that it was created.  If you edit a drawing, the stuff you edit, no matter where in the drawing it is located, is last on the instruction list.  AutoCAD is a fine for tool for creating and printing a drawing, but the text that instructs the drawing isn't very good to give instructions to a human.  Once the entire drawing is drawn, it can be interpreted by a human, but the machine language is only for machines.

Pat

roo

  • Sr. Member
  • *****
  • Posts: 1855
    • The Notable Knot Index
Re: Houston - We Have a Problem
« Reply #17 on: December 12, 2006, 04:32:40 AM »
...
Most of us can draw with a pen and paper, but have you ever tried to draw with "Paint" ?  No matter how careful you are the result looks like a 2 year old did it.  Consequently we all try to get by with words rather than expose the world to our Painter works of 'art' and ourselves to rigorous jesting.
...
Derek


Don't worry, nobody will give you (or anybody) a bad time about sloppy or choppy drawings.  Most of us are relieved and grateful to get an unambigous visualization we can decipher in a few seconds... and on the first try.   ;)
If you wish to add a troll to your ignore list, click "Profile" then "Buddies/Ignore List".


admin

  • Administrator
  • Full Member
  • *****
  • Posts: 77
    • Black Widow Web Design Ltd.
Re: Houston - We Have a Problem
« Reply #18 on: December 31, 2006, 09:47:51 PM »
We will always need pictures, but I believe we also need a simple and unambiguous text based system that we can fall back on when needed, we just need to agree on one.

Speaking as a programmer and web developer, I can appreciate your point. You're looking for an information communication system specifically for knots, yes? How that information is interpreted when it reaches its destination is almost a separate issue. Once the 'data' is received, it can be rendered in a number of ways that are appropriate to the person/system receiving it. The key issue, as I see it, is ensuring that the 'data' moves from person to person (or system to system) in a standard, unambigious manner.

I'd suggest you look at XML - eXtensible Markup Language. This is a system that is designed to communicate information but, most importantly of all, you can define your own 'language terms'. Your own 'dictionary' if you like. As long as people/systems have access to your 'dictionary', they can understand the information that they have been sent and can 'decode it' correctly according to their needs.

http://www.xml.com/pub/a/98/10/guide0.html

However, it is my experience that many of the people who excel at arts/craft are those who find logic-based solutions to be the most alien. It seems to be a different mind-set. So it may be that some of the very people who could make good use of your Knot Markup Language will find it uncomprehensible and/or intimidating. That said, it would seem to be an excellent possibility for archiving material in a form that would be (relatively) easy to extract at a later date.


KnotMe

  • Sr. Member
  • *****
  • Posts: 596
    • The Dao of Silk
Re: Houston - We Have a Problem
« Reply #19 on: January 02, 2007, 01:16:39 AM »
Is not Adobe Illustrator (or other vector graphics drawing programs like Corel Draw or Inkscape or Freehand) a good compromise between AutoCAD and raster drawing programs like Paint (or Photoshop or Gimp or Painter or...)?  If 3D is needed, there are a slew of other open source programs, plus IIRC Maya is free to use for personal purposes.

A vector drawing programs manipulates line segments and curves vs pixels producing a file format that can be more easily modified and better approximating the representation of cord.

Inkscape (and probably others) uses SVG (XML based representation of the vector information).  Most knots, at least during the tying, seem like they can be represented in 2D given some layer information.  So, using an SVG based vector drawing program would give you the tedious computer representation in a relatively painless fashion.  In 3D there is OpenGL among other formats.

If we're just talking about helping our fellow knotters tie a particular knot, then clear illustrations and some explanatory text is the basic thing to do.  You can add in videos of the construction process.  The next stage is probably a 3D animated model that can be rotated.  After that, maybe you can give people kits for each knot with a couple of tied knots so they can inspect them and move the cords around.

Anything that completely and unambiguously describes all cases of knot tying is bound to be extremely technical and arcane.

What I'd love to be able to do is come up with a computer representation system that spans Celtic knot drafting (several of which already exist), turk's head drafting (ask the system for an AxB turk's head and get the grid, there's at least 1 or 2 of these, right?) as well as, say, AxB mystic knots or A over B flower knots or AxB bao knots or ... etc.

It's possible that a representation system and a construction description methodology can be generated by the IGKT big brains, academics and theorists, but it also might be quicker and easier to do it on a case by case basis (plug-ins!): such as expandable rectangular turk's heads, expandable cruciform turk's heads, solid convex mat shapes, etc.
« Last Edit: January 02, 2007, 01:55:18 AM by KnotMe »

DerekSmith

  • IGKT Member
  • Sr. Member
  • *****
  • Posts: 1521
  • Knot Botherer
    • ALbion Alliance
Re: Houston - We Have a Problem
« Reply #20 on: January 02, 2007, 05:15:36 PM »

Anything that completely and unambiguously describes all cases of knot tying is bound to be extremely technical and arcane.


Arcane is an interesting word.

Dict Arcane  known or understood by very few; mysterious; secret; obscure; requiring secret or mysterious knowledge.

When I first applied for apprenticeship to the world of Knot Tyers, I was introduced to the words of the Guild - Bend, Bight, SPart, WorkingPart etc. (known or understood by very few; mysterious; secret; obscure; requiring secret or mysterious knowledge.)  I think we are struggling because the knotting world is already 'Arcane' - and worse - it is inconsistent.  Ashley defines a Bight as the centre part of a rope or a curve or arc in the rope, while many will take the term to mean an open loop (ABOK #31) or a closed loop (ABOK #32).  Layer on top of this the diversity of terminologies stemming from different trades and you have the near useless lexicon with which we struggle (and generally fail) to communicate today.  These words have worked for centuries because in the day of the Apprentice, the fundamental means of transferring knowledge was demonstration.  Words were to help recall, but the essence of knowledge transfer was done at the hand of the Master by visual example and demonstration.

Today, we do not sit at the Masters side as they demonstrate their knowledge.  We sit on opposite sides of the globe and struggle to use the words in isolation.  It should come as no surprise that these archaic words fail us, it should not be too much of a surprise to learn that these words MUST fail us - they were never designed to be the primary means of teaching (knowledge transfer) and they fail the task miserably.

BUT - today, we share more knowledge around the world in a second, than would have been shared in a lifetime when our knotting lexicon was created, and it is nearly all done in words.  Special words, designed to transfer knowledge to others around the globe, without fear of ambiguity.  These words are the language of HTML and XML and you may well choose to describe them as "extremely technical and arcane".  In a small way you would be right, but you would be much more correct to describe them as "precise and simple".  You could learn to write (and read) a simple html page in under an hour, in contrast, how long would it take you to learn the simplest of  spoken languages with its book sized dictionary?  The Internet has taught us that to exchange information unambiguously, we need fewer words, not more and a simple syntax that does not change meanings with the experience and perspective of the reader.  The only downside of this is that it takes a lot of simple descriptions to describe something which is complex, but that is a tiny price to pay for a communication system which actually works.

Although HTML and XML are designed to be machine readable - i.e. a computer can 'read' them and create on the screen the information intended - XML also has the fundamental prerequisite that it be human readable (and of course, writeable) it might be tedious in its simplicity, but it is 100% accurate in its reproducibility and 0%  ambiguous in its interpretation and that surely is what we need to start off with in our struggle to create a knotting lexicon for the 21st centuary.

Quote
What I'd love to be able to do is come up with a computer representation system that spans Celtic knot drafting (several of which already exist), turk's head drafting (ask the system for an AxB turk's head and get the grid, there's at least 1 or 2 of these, right?) as well as, say, AxB mystic knots or A over B flower knots or AxB bao knots or ... etc.

It's possible that a representation system and a construction description methodology can be generated by the IGKT big brains, academics and theorists, but it also might be quicker and easier to do it on a case by case basis (plug-ins!): such as expandable rectangular turk's heads, expandable cruciform turk's heads, solid convex mat shapes, etc.

I think that the goals you expound for yourself are way, way ahead of the 'simple' knot description language I am seeking - on the scale of 'walk before you run' you are probably just boarding the Mars Express!!  However, when we have a language that can reliably describe a simple knot, then we are only a step away from describing complex or compound knots.  Then in turn we are only a step away from describing tiles or to use your phrase 'plug-ins' and that in turn is only a step away from being able to assemble the highly complex constructions you are used to working with every day.

Every great journey starts with the first step, and perhaps if we can build on Mel's suggestions of basing our lexicon on XML, then maybe we have taken our first few faltering steps.  To proceed, some of us are going to have to learn a little XML and create and define a handful of terms that will allow us to 'write down a knot'.  We will then have to apply this to a selected handful of benchmark knots to uncover 'special' cases, and develop methods to handle them.  Key to this phase will be uncovering all the questions like - "How do I handle objects included in the knot like - eyes, spars etc.?" and "How do I annotate a helical coil?".  While some of us have our noses proverbially at the coal face, we may well need others (maybe Mel?) looking at the bigger picture in terms of how we go about integrating our little bit of XML with the mainstream and how we go about creating an XML parser that can render our  descriptions into glorious 3D in our browsers, and maybe even the reverse - taking our sketches or pictures and turning them into XML for us?

It will be quite a journey, but I am confident it can be done and equally certain that we need it now we no longer have apprenticeships to Master Knotters.  We will of course need a place to collect the ideas and developments together and as ever there is a wiki just waiting to oblige at http://knot-html.pbwiki.com/  Any comments, corrections ideas etc can be entered using the password igkt backed up with discussion through the forum.  If you have already spotted some problems/hurdles then jot them down on the problems page at http://knot-html.pbwiki.com/Problems%20and%20Challenges someone is bound to have ideas how to approach them.


All Aboard for the Mars Express ??

DerekSmith

  • IGKT Member
  • Sr. Member
  • *****
  • Posts: 1521
  • Knot Botherer
    • ALbion Alliance
Re: Houston - We Have a Problem
« Reply #21 on: January 02, 2007, 06:29:57 PM »
We will always need pictures, but I believe we also need a simple and unambiguous text based system that we can fall back on when needed, we just need to agree on one.

Speaking as a programmer and web developer, I can appreciate your point. You're looking for an information communication system specifically for knots, yes? How that information is interpreted when it reaches its destination is almost a separate issue. Once the 'data' is received, it can be rendered in a number of ways that are appropriate to the person/system receiving it. The key issue, as I see it, is ensuring that the 'data' moves from person to person (or system to system) in a standard, unambiguous manner.

I'd suggest you look at XML - eXtensible Markup Language. This is a system that is designed to communicate information but, most importantly of all, you can define your own 'language terms'. Your own 'dictionary' if you like. As long as people/systems have access to your 'dictionary', they can understand the information that they have been sent and can 'decode it' correctly according to their needs.

http://www.xml.com/pub/a/98/10/guide0.html

However, it is my experience that many of the people who excel at arts/craft are those who find logic-based solutions to be the most alien. It seems to be a different mind-set. So it may be that some of the very people who could make good use of your Knot Markup Language will find it incomprehensible and/or intimidating. That said, it would seem to be an excellent possibility for archiving material in a form that would be (relatively) easy to extract at a later date.



Mel,

As ever, I find your analysis to be spot on.  You focus precisely on the requirement of creating the knot description in "a standard, unambiguous manner".

Your suggestion to use XML is probably key and I am reading through your link pages to get an understanding of the XML requirements, of course, any tips and help you can give me would be appreciated.

I think you will be particularly correct that art/craft folk will not warm to the pedanticity of writing a knot description in XML or any other form of  'RISK' language.  Consequently I think we should not expect too much participation from the knotting membership, at least until the thing starts to bear fruit and we start to see the descriptions being rendered on screen.

I particularly relate to the need to archive knot information.  Pictures and written descriptions are often insufficient to understand the whole structure of a knot and diagrams fail to define the dressed or working structure of a knot.  Even Ashley 'lost' some knot definitions and I am sure that knots are being continuously lost as people die taking with them their unrecorded memories.  A Knot-XML will mean we can record the structures of knots in their dressed structure and will give the additional advantage of allowing us to search for 'signatures' of various knot elements within knots - this has obvious forensic values and will allow knots to be studied as family groups sharing specific structural characteristics.  I remain firmly convinced that the future development of any 'Science of Knots' depends firmly on the prior development of a knotting descriptive lexicon - a 'language of the knots'.

I only hope the project attracts enough people to make a job of it.

KnotMe

  • Sr. Member
  • *****
  • Posts: 596
    • The Dao of Silk
Re: Houston - We Have a Problem
« Reply #22 on: January 02, 2007, 08:29:59 PM »
I think that the goals you expound for yourself are way, way ahead of the 'simple' knot description language I am seeking - on the scale of 'walk before you run' you are probably just boarding the Mars Express!!

I always do that.  I take one or two baby steps then tend to head straight for the Mars Express.  8)

I'm just going to say one more word about XML and then shut up, though.

It's true that one of the requirements of XML is that it must be human readable.  This is to prevent it being co-opted by powers that may wish to corrupt it and try to take it into the proprietary realm (as has happened to HTML in the past and CSS in the present).  But the primary purpose of XML is to describe data for interchange via computers.  Tagging parts of a book "author's dedication" "chapter head" or "table of contents" is for the benefit and understanding (data parsing and extraction) of computers, not humans directly.

To create a new XML schema, a human must first thoroughly understand the topic and then create a breakdown for computer input.  Then they need to create the software that will decode it again for human understanding.  In other words you need to know how to take your baby steps and also have a complete map to Mars and Alpha Centauri before starting to share or risk littering the landscape with droves of incomplete and possibly incompatible "one true description language".

If your primary purpose is to transmit data from human to human then sitting down as the IGKT and agreeing on bight vs open loop, working end vs bitter end vs etc. and standardizing (and clearly, formally defining) would go further, more efficiently.

DerekSmith

  • IGKT Member
  • Sr. Member
  • *****
  • Posts: 1521
  • Knot Botherer
    • ALbion Alliance
Re: Houston - We Have a Problem
« Reply #23 on: January 02, 2007, 10:18:40 PM »
Wow,

Carol, am I glad you are here, You clearly know a lot about XML and its use and perhaps more importantly, its issues.

Please do not make this your last word on XML, I feel that if I can hang on your coat tails long enough, your knowledge and energy might just help take me where I need to get.  There is much I need to learn, not least is the very point you make that in order to create a language to describe a knot, we must first learn how to describe it.  Doubtless, with that process there will be mistakes and dead ends, but I do not see a new asteroid belt of dumped dictionaries of failed attempts.  More likely, the language will acquire necessary components while discarding those which were inadequate - it will evolve into something usable, and if the cost of that is a few dead end remnants, then that is a small price to pay.

Yes, I understand that the purpose of XML is for it to be machine readable, and yes that would have to be seen as an ultimate goal, but along the way we have first to learn, as you carefully have pointed out, how to describe a knot, and in doing so we will uncover the means of unambiguously communicating knot movements/structures to one another.

As for your suggestion that we should consider sitting down as the IGKT and resolving the disparity of perceptions ranging over our existing terminology - well - first, if it could be made to happen, it would be a welcome breath of clarity to our field - BUT -

It would not resolve this problem, because the terms themselves do not, and cannot, describe the subtle angles and twists which differentiate one knot from a close neighbour, and because you would still be using unrestricted English, you would not be removing the opportunity for ambiguity from a description.

However, the really big BUT is - how long are you planning to live for?  From what I have seen of the IGKT, the only good things to happen are this Forum and the KM - Nothing else Happens, or at least if it does the pace is measured in lifetimes.  I genuinely think that without being naive about this, the only way to make changes happen is for members to do 'it', then hope that the IGKT PTB get round to ratifying it (but don't hold your breath).

Besides, I could not sit down as the IGKT, because I am not a member yet - my application made nearly 8 months ago is still somewhere in the system and I have only ever had one reply from any of my enquiries re progress (like I said, progress measured in lifetimes and Stasis boxes haven't been invented yet!).

Knot Head

  • Knot Head
  • Exp. Member
  • ****
  • Posts: 194
  • A knot is not just a knot...
Re: Houston - We Have a Problem
« Reply #24 on: January 03, 2007, 05:28:25 AM »
Hi all. Forgive my ingnorance in this matter. I have found that a picture, or even an illustration can replace a thousand words. But if you combine the two in one place you have an indespensable resource of information to learn and to be taught.
Regards,
Brian Kidd