Multiple simultaneous galleys
This is a modal window.
Das Video konnte nicht geladen werden, da entweder ein Server- oder Netzwerkfehler auftrat oder das Format nicht unterstützt wird.
Formale Metadaten
| Titel |
| |
| Untertitel |
| |
| Serientitel | ||
| Teil | 19 | |
| Anzahl der Teile | 33 | |
| Autor | ||
| Lizenz | CC-Namensnennung 3.0 Unported: Sie dürfen das Werk bzw. den Inhalt zu jedem legalen Zweck nutzen, verändern und in unveränderter oder veränderter Form vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen. | |
| Identifikatoren | 10.5446/30794 (DOI) | |
| Herausgeber | ||
| Erscheinungsjahr | ||
| Sprache | ||
| Produktionsort | Cork, Ireland |
Inhaltliche Metadaten
| Fachgebiet | ||
| Genre | ||
| Abstract |
|
1
2
4
7
8
12
17
18
19
20
24
28
30
00:00
Quick-SortFlächeninhaltAlgorithmusInformationsspeicherungTouchscreenDatenstrukturPunktVisualisierungSpannweite <Stochastik>SelbstrepräsentationDifferenteZeichenketteFunktion <Mathematik>Wort <Informatik>BitDivergente ReiheSynchronisierungRandverteilungOrdinalzahlGarbentheorieFunktionalArithmetisches MittelGruppenoperationGraphfärbungWeb-SeiteEndliche ModelltheorieAuszeichnungsspracheGrenzschichtablösungWurzel <Mathematik>Deskriptive StatistikIndexberechnungMultiplikationsoperatorBildschirmmaskeFormale SpracheMailing-ListeTypentheorieGeradeZahlenbereichÄquivalenzklasseTranslation <Mathematik>Paralleler AlgorithmusMultiplikationTupelAutomatische IndexierungMechanismus-Design-TheorieKontextbezogenes SystemRadikal <Mathematik>Ein-AusgabeStreaming <Kommunikationstechnik>SkriptspracheMereologieCASE <Informatik>InstantiierungMinkowski-MetrikFolge <Mathematik>OrtsoperatorTransformation <Mathematik>TabelleRechter WinkelGesetz <Physik>ComputerspielVersionsverwaltungVorzeichen <Mathematik>Natürliche ZahlAutomatische HandlungsplanungSummierbarkeitMatchingKlasse <Mathematik>FrequenzStichprobenumfangSchnittmengeGrundraumRobotikPhysikalische TheorieVollständiger VerbandInhalt <Mathematik>AggregatzustandVideokonferenzBildgebendes VerfahrenObjekt <Kategorie>BildverstehenMetropolitan area networkComputeranimation
Transkript: English(automatisch erzeugt)
00:04
Okay, so as John said, he was talking about individual, if you wish, atoms, and he presented a whole series of different atoms. He presented characters, and he presented words, come from words.
00:21
So here I made a summary of a few of those items that we could think a document can be represented by. So we have, so the character we believe is stem. Prefected suffixes, declensions, word, come from word, a phrase,
00:42
a sentence, section of a pattern, and so on. You can imagine that you could have something else in between. You could have combinations. You can have just a group of characters that supposedly for somebody to mean anything, but they might have another meaning for somebody else. So to sort of take him from John, we have atoms.
01:01
So what do we do with them? So what we're really proposing is a sort of a document model that will have parts to it. And so the first one, the easier to think about is a sequence of these atoms, and that's what we call a galley.
01:22
But one galley on its own, and it's not very interesting. So we have thought of a mechanism of joining these galleys together, and that's what we call a list of links or a link, simply. You can think of those algorithms as a manipulated galley.
01:40
Transforming one into another, if we have a list of characters, we can produce a list of words. Or if we have a list of words, can we provide a list of sentences? So this is parsing from input to galley.
02:00
So if we have an editor, I want to be able to take that input and put it into a galley that I want to use for a specific algorithm that you want. And then obviously taking it out from that storage and put it into an output form of whatever I decide my output form should be.
02:20
And obviously input and output mechanisms. So let's talk a little bit about galleys and how do we might think of them. So here I just have a very, you know, simple sort of string of characters, and we actually have here, we can see two galleys.
02:43
If you first look at the, my left side is the WA1, WA2. So we can think of a galley that has words and they're indexed by WA1, WA2, WA3.
03:03
To the right side, we can have another galley that is just representing characters. Okay? And yeah, for now, just think of them like having just these two different representations. So an example link will be what?
03:23
So if we go back, we have the last bit of the word primary from 228 to 232. So we can just sort of grab that range and say, I want to have all this range linked to a function g of c
03:46
where I can say that g of c is color blue. So this is what we can say is, in other words, markup. I'm marking up. That section of my galley to be printed in blue in my output. Okay? But you can imagine that that function can be anything.
04:09
Another example of multiple galleys, and you can imagine that this would have any script that just wasn't quite easy at that moment, we have three separate word galleys
04:22
in three different languages. And arbitrarily, we have indices for each of the words. So this came about when we were thinking that how could we easily think of printing parallel text in different languages?
04:40
So how are we going to be able to match? Okay, I want both without having to be dealing with translation. So well, I have this first paragraph, and I want this whole paragraph to match in Spanish with this political paragraph in another language, and this political paragraph in another language. So having this list of words which have indexes for each
05:02
of the words, then we just create a tuple in which we say that first tuple, the beginning of that tuple, the text region should match with the beginning of that one and the beginning of that one and the same before the end. And you can imagine the algorithm being able to say, okay,
05:21
I know that I have typeset these three chunks of text next to each other in a specific area of my visual extract. And this is just a supposed output that we can imagine that up to this point of that part,
05:41
all the three paragraphs are connected at that point. And that would apply, especially if you have a screen that you could imagine having the things that are next to each other and that they are matching visually. So that brings us to something
06:01
when what happens we have several galleys. In the previous example, all the galleys were sort of at the same level, but what happens when we have the main galley of being just text, but we also have another galley which has the footnotes, and we also have another galley that has the markup like the functions of color that I mentioned before.
06:23
So this has to be interleaving. We have to find mechanisms to be able to point from one to another. The reason why we put them separate is because we want to be able to say usually marginal notes and critical notes are not considered
06:41
to be part of the main case. If you're doing some sort of search, you want to be able to differentiate it too. So we could also have tables the way John mentioned them, that they could just be a tuple, and that could be another galley
07:01
that gets inserted somewhere. So now we have this point just to the main galley which John talked about being the iterators, and that's how we would manipulate the sort of ranges. So what are these lists of links? Well, they would be represented by tuples.
07:24
They could be a point or a region, and they have, since we could have separate lists of these links, they can be overlapped, and they can be nested or overlapped,
07:41
so that's probably what John was thinking about for your question about music, you can have one galley representing one sort of linear, you know, in time, and then having another one which overlaps in another time frame, if you wish.
08:01
You have that marker equivalence between galleys, which is what we're doing with the paragraphs. So that's many things. So this is an example about a very simple word, a happy word for mammals, and we have how we can have three different points
08:22
and representing three different places in the word. So if we have a pointer, if we have a range 3.3, that means that that's the position between the roots and the termination of the word. And so all we're seeing is just a root, and that's just to show
08:43
that we can represent absolutely any point in this stream of characters, any point or any region that we wish to, and we can often, we can tag it in our galley. We have just roots or something like that. We can just say, okay, this is one, this is another one,
09:01
and have specific descriptions for each of them. So, well, this is just really much brainstorming what sort of algorithms we can think of applying to this galley. Well, you know, this is an infinite list. We can do, we could imagine doing any of this at any time.
09:27
Line numbers or types of text, for example. We can have one galley and convert it into another galley. Panel text, we can imagine being able to use a very flexible range of algorithms.
09:43
So, and the output is the same with the output given, you know, just the algorithm. What is really, really, really important here is that the storage of the galley, the galley, the data structure of the galley be flexible enough for the algorithms to be able to manipulate them
10:01
and to be able to manipulate new galleys as is needed. And those are the end comments of the model. This is just, it would be so much flexible that anything that we've pretty much been seeing.
10:23
And the key point is really the design of these data structures in the galley mode, having this link, atoms representing different things of a text. And that's it. Any questions, any comments?
10:42
Yeah. When you speak about synchronization between galleys, it obviously means different things for different kinds of galleys. For example, if you have your individual strings, your synchronization means temporal in the same time. But when we speak about typesetting, it's completely different things. It means maybe on the same places
11:01
on different pages or in the different. So you, I think that in your model, you need to have some, besides the general term, you need to have some, you need some instances. What do we mean by synchronization for any kind of data, right? Yes, this is, at this point, it's more of an abstract model
11:22
that once you start going down into the specific system, we start applying what you just said. Yes, this will be just time. Yes, this will just be position. Synchronization in position. Yes, exactly. My impression of what I've said
11:41
is that there are different kinds of gals, those which can be derived by transformation from others and you have those which you absolutely have to keep because persistently because they are, they are essential to keeping the semantics of everything.
12:02
So probably there should be kind of classification or something to indicate. I think that that's what applies to what John said about certain representations who have a canonical representation of their text and that would be their base. And everything would be the back of them,
12:20
but somebody else might have another base that they want to keep as their canonical that you just keep all the semantics, yeah. Wonderful. I have to be very careful to throw away representations. I have seen that in some research projects, people have thrown away representations on how it was put into old books
12:41
and they were credited dearly afterwards. You see how far the margins were bad. These simple things as how far, how wide the margins were was actually important for cultural research. So what you're saying is with any form of input into the system,
13:01
we have to be out of it. In the end, yes, yeah. Yeah, no more. You don't know what kind of questions people are going to be to ask on your stuff. So there's something that you might see as something derived and you don't care for it.
13:21
And suddenly some historical research person comes across and wants to research but how far into what space it is because you're doing research on typographical traditions. And suddenly he's interested in, well, here we have narrow typesetting
13:40
and here we have wide typesetting and how it's changed over the years. And so normally you would say a blank or a wide space as to just turn on the, not just get the discussion between you and Jonathan or it's not relevant but it might be relevant in the context of research
14:02
of traditional typography. Any other questions?
Empfehlungen
Serie mit 1 Medium