Version classiqueVersion mobile
OpenEdition Books

Património Textual e Humanidades Digitais

 | 
Maria Filomena Gonçalves
, 
Ana Paula Banza

Easily annotating manuscripts online

A Web-Application for linking images to texts

Manuel Raaf

Texte intégral

1. INTRODUCTION

  • 1 http://www.proloewe.de/en/home/welcome.html
    http://verwaltung.hessen.de/irj/HMWK_Internet?uid=fa560c (...)

1In May 2012, the development of a linguistic tool for annotating palimpsests started at the LOEWE-Center “Digital Humanities” and its subproject Caucasian Palimpsests, housed at Goethe-University Frankfurt, Germany.1 The LOEWE-Center is funded by the State of Hesse and contains several projects that make use of technologies as tools for academic and scientific research. As part of the focus on “Digital Humanities”, researchers of the Humanities and computer scientists combine their skills in creating (large) text-based corpora and tools for using and expanding them.

2Work began in June 2012 on the annotation software. During the first month of preparation, different possible techniques for realizing the aims described in chapter two were considered. Significant progress has since been made; however, since creating the application is a one-person, part-time position, there is still a lot to do. Nonetheless, the aims given next will definitely be reached at the end of the project’s runtime in December 2013, and the application can already be used in its current state for basic annotations.

3Below is a description of the project’s aims and the current state of the application, as well as plans for the future and a short summary about the work done so far. There is currently no public demo available. For technical details, see chapter 4.

2. AIMS

2.1. GENERAL AIMS

  • 2 Unicode is a standard for encoding and displaying characters and symbols in most of the world’s lan (...)

4Although the project title expresses a relation to Caucasian scripts, the application need not be restricted to this project only. Rather, it has been designed such that any type of character or symbol that is supported by Unicode2 can be entered. The number of projects supported by the application, as well as the number of annotations in one project, is similarly not restricted. Furthermore, the tool is able to satisfy linguistic needs in searching for words, word forms, allomorphs, etc. Additionally, all this is provided online in a user-friendly way, without the need for installing any software or browser-plugin.

5The kind of images that can be annotated by means of the application can be chosen freely by the user, so that manuscripts, palimpsests, inscriptions, graveyards, music notes, and many more types of images can be annotated.

2.2. REASONS FOR DEVELOPMENT

  • 3 Modern hard disks in workstation computers (e. g. office computers) usually read around at least 60 (...)

6Computers are a very powerful tool for linguists (and other researchers of the Humanities). By means of their speed in reading and processing data,3 one is able to search through thousands of pages or even annotate texts automatically, thereby saving one a lot of time compared to reading, analyzing and comparing the material by hand. With a bit of training, any researcher will find computers relatively easy to use.

  • 4 XML (Extensible Markup Language) is a markup language that is used mainly for hierarchically saving (...)

7Currently, existing annotation software is limited by the fact that it has to be installed on a computer and does not allow for annotating multi-layer images, e. g. palimpsests. It is usually also not yet possible to add more images to one set of annotations. In some cases, it is also not possible to enter data without the knowledge of markup languages, such as XML.4 Hence, the software presented in this paper is not merely reinventing the wheel, but rather the necessary creation of a new kind of annotation program that solves the aforementioned weaknesses.

  • 5 I recommend Mozilla Firefox (http://www.mozilla.org) for several reasons including modernness, secu (...)

8In addition to the general advantages of computers as linguistic tools mentioned briefly above, there are several more specific benefits offered by web-applications. The most important of which is the lack of need to install any additional software on the researcher’s computer. Consequently, making and reviewing annotations is not bound to one specific computer, rather one is able to access and use the application and all data from any computer that has access to the Internet. The only software requirement is that the user needs a modern browser.5 The computers and notebooks used by scientists today are usually up-todate sufficiently.

9By storing all the data online, one is able to access it from any computer from anywhere in the world. There is no need for manual backups, since all the data is stored on the server and backups will be performed automatically by it. The computing power of a server is, on average, much higher than that of a workstation (i. e. personal computer, notebook). Therefore, the application is faster in processing, e. g. search queries.

10Thanks to the Internet, one is also able to share work with colleagues from all over the world, as well as to publish work or research questions.

3. THE APPLICATION

11This section describes in brief the application’s parts and functions in their current state. Except two of the application’s main parts for annotating images (see 3.4), screenshots of the applications will not be presented due to limited space in this paper.

12The application’s current working title is ImAnTo (Image Annotation Tool).

3.1. PROJECT MANAGEMENT

13Since ImAnTo is able to handle more than one project, there is the need for a project management system. This allows users to create new projects, as well as to change or delete existing ones. The basic information for each project is currently limited to the name of the project, its description, the default sublines, and default number of fields per line. The latter settings save time when annotating because each subline containing the given number of fields will be created automatically for each master-line (i. e. the line containing the text visible in the image). In other words, standard sublines such as transliteration, glossing, translation and comments can be created by default for each added master-line. After creating a new project, one is able to upload images.

3.2. UPLOADING IMAGES

14Storing images to a project is possible by uploading either single image files individually or as an archive (e. g. zip, tar, 7z, etc.) containing any number of image files. ImAnTo will automatically create all preview images upon upload. Since all images are stored to one specific project, the set of all images can be edited via the Image Browser.

3.3. IMAGE BROWSER

15Whenever multiple images are assigned to a project, a comfortable way for browsing through them is provided: the image browser. This displays small preview-images (so-called thumbnails) in a list similar to the image preview in a folder on a Windows-system. Each of the thumbnails is followed by a checkbox that can be checked in order to delete the chosen image (see Figure 1). Additionally, each image can be enlarged by simply clicking on it — doing so will provide a closer view (see Figure 2). Next to the enlarged version are several links: one to the Annotation Page (see 3.4.) for this particular image, one to the Setup Page, and one for deleting the image and its associated annotations. Clicking either the first link or the image itself opens the Annotation Page. The Setup Page provides basic functions for editing (e. g. cropping, rotation), deleting, adding comments to, and adding alternative versions of the image. Alternative versions might be needed in cases where multispectral imaging was used for creating the images. Hence, since those images are in grey-scale, the original/colored version could be added as alternative image, to which one can switch later on the Annotation Page.

Figure 1 – Thumbnails of all images of the selected project

Figure 1 – Thumbnails of all images of the selected project

Figure 2 – Preview of the selected image

Figure 2 – Preview of the selected image

16Each preview can be easily closed (similar to closing programs or folders in Windows) in order to go back to the complete list of the project’s images.

3.4. ANNOTATION

3.4.1. DRAWING SHAPES

17The important information in the image has to be highlighted in some way in order to link it to the particular textual information. For doing this, ImAnTo currently provides three different shapes: rectangle, circle, and polygon. Each can be easily created by selecting the desired shape, then clicking and moving the mouse to create the shape nodes. Once drawn, the shape can be moved, resized and/or deleted by using the mouse and/or the keyboard, similar to the use of picture editing software such as Windows Paint or PhotoShop.

18The shapes are stored in a special object within ImAnTo, i. e., they are not really drawn on the image. Consequently they do not damage/modify the image, rather simply overlay it.

19Each new shape will automatically be linked sequentially to the first column of text fields that is not yet linked to a shape, e. g., the first shape to the first column, the second to the second, the third to the third, and so on. However, this relationship can be set by hand (using mouse and keyboard combinations) as well.

20A future update to ImAnTo will allow for linking more shapes to one text column, i. e., so that researchers are able to group shapes for word-wrapped text, etc.

3.4.2. TEXT FIELDS

21All textual information, e. g. the annotation data, is stored in text fields and encoded in Unicode characters. ImAnTo allows researchers to add as many sublines of annotation as required, so that one is able to add sublines for glossing, translation, transliteration, comments, alternative translations, POS, etc. The structure of lines and their fields resembles a spreadsheet table, in which an entire column of fields within a given master-line is linked to one shape on the image (see Figure 3). All sublines will be added below the master-line. By clicking the button (or the menu entry) for “new line”, “new field”, “delete line”, or “delete field”, the associated action will be executed.

Figure 3 – Screenshot showing part of the Annotation Page

Figure 3 – Screenshot showing part of the Annotation Page

3.4.3.. ANNOTATION LEVELS

22Annotations represented in one master-line and as many sublines as needed will satisfy most users. Each text field (at a minimum the ones in the master-line) will usually contain one word. However, manuscripts often contain letters of various sizes or in a different color, e. g. drop caps, etc. In order to annotate this information, too, the content of the text field needs to be split into another subline, i. e. another annotation level itself containing new sublines (see Figure 4). This can easily be done by clicking on “add character level”. ImAnTo will automatically split the field’s content into single letters, each of which appears in the first field of a new text column of the new level. All other levels will be faded out in order to focus solely on the selected one. For example, the sublines related to this new level might be named description and colored, containing information like “capital Latin letter c” and “red”. Searching for these additional levels of annotation is possible by selecting them individually on the Search Page (see 3.5). It is possible to add as many levels as needed, each of which contains as many lines as needed. Therefore, it is conceivable even to split single characters further into sublevels, if needed, in order to annotate strokes or other atomic characteristics, e. g. for hieroglyphics or Asian scripts, etc.

23Figure 4 shows an example of this feature: after clicking either the small button above every text field or the menu entry “add character level”, the field’s content will automatically be split into its characters. In the example below, the sublines are named character and description. In the case that any of these new fields do not need additional annotations, they can be left empty. Due to space limitations in this paper, only a small section of the Annotation Page can be presented in the following screenshot.

Figure 4 – Screenshot showing part of an annotation; focused on character-based first sublevel

Figure 4 – Screenshot showing part of an annotation; focused on character-based first sublevel

3.4.4. ADDING ADDITIONAL IMAGES TO ONE ANNOTATION

24The challenge for annotation software for palimpsests lies in the nature of palimpsests themselves, i. e., that at least two texts and/or scripts overlay one another. A particular image of the palimpsest may show most of the underlying script very well, while some parts can only be visible in another image. In order to annotate the whole palimpsest as a single item, there hence need to be additional images available to the annotator.

25To do so in ImAnTo, each shape and its related text column can be associated with a second image taken from the images uploaded to the project. The user simply uses the function for adding another image that then can be selected in the image browser. This will then be loaded in a popup-window (i. e., one does not have to leave the page for annotation). Afterwards, ImAnTo links the additional image to the particular text column. By selecting either the related shape or the column itself, the image will be made visible automatically.

3.4.5. GROUPING SUBLINES

26If desired, researchers can group sublines in order to hide unneeded ones. For doing so, the groups need to be defined in the Project Manager. The Annotation Page will then contain links for showing/hiding these groups of sublines. This will help when dozens of sublines for each master-line are used, because they can be temporarily hidden in order to focus on those that are relevant. For example, a manuscript containing 30 master-lines, each of which contains on average 7 fields by 50 sublines, would have a total of 10,500 fields — this would lead to a very long Annotation Page. Being able to minimize unneeded sublines reduces both screen space and potential user confusion.

3.4.6. UNDO, REDO

27Since one of the stated aims is usability, ImAnTo provides a function for undoing and re-doing an action — moving, adding or deleting a shape, as well as changing the content of a text field or deleting it can be easily undone. If necessary, a given action can be “re-done”, i. e., the process of undoing can itself be undone, thereby restoring a previous change.

3.4.7. RE-INDEXING

28The removal of one text field will, of course, cause the deletion of the whole column, but it will not delete the shape to which it is linked. After deleting several text columns, the abandoned shapes will definitely need to be linked to new annotation columns. In order to do so, one simply uses the function for re-indexing (i. e. re-linking) the shapes to the columns hierarchically according to the order of the text columns: the first shape drawn on the image will be linked to the first text column, the second shape to the second text column, and so on. Unfortunately, it is not possible at the moment to use the position of the shapes on the image as the basis for re-indexing the relationships. This further development might be implemented in the upcoming months.

3.5. SEARCH

29The current search function allows users to select a project in which to search, as well as its existing lines in order to focus on particular annotations, e. g. in the gloss-line. The wildcard * stands for “any character from zero occurrences to infinity” and allows one to search for word forms or morphemes, e. g., * ção as search value will provide introdução, as well as imaginação, etc. The results will be listed either as pure text (see Figure 5) or as text extended with its associated shape in the image in which it occurs, so that one is able to quickly see the particular part of the image that belongs to the search result.

30In addition, each result is highlighted and by clicking on it, a new window containing the particular Annotation Page opens. There, the shape associated with the particular search result appears as the only one in order to focus on the result. All other shapes, which are temporarily invisible, can be made visible again by choosing the option “show all shapes”.

Figure 5 – Search Page with highlighted results

Figure 5 – Search Page with highlighted results

3.6. ZOOM

31In general, since about half of the screen will be used for displaying an image, it should be big enough to be readable. Still, one might want or need to zoom into specific areas in the image. By clicking the “zoom” button and moving the mouse over the image, that part of the image will be magnified at the coordinates of the mouse pointer.

4. TECHNIQUE

32This section describes the application’s technical aspects in brief. More details about the single techniques mentioned below are provided in the particular footnotes.

  • 6 Open source software is free of charge and anyone is allowed to create his own version by modifying (...)

33All languages and programs are open source, so that there is no need for any licenses. There are well documented instructions, manuals, help, performance and security statistics available for all.6

  • 7 Apache is a very popular and powerful free web server that enables one to involve database language (...)

34As web server, Apache7 is used for providing and processing the data. It is locally installed on a standard office-computer that serves as client and server. The system has an Intel i5-660 CPU with 3.33 GHZ and 4 GB of RAM. As operating system, Windows 7 Professional 64bit is used. There was no need for buying a special computer as a development system and/or dedicated (test-) server, since the application does not need much computing power in its current state and usage. The client (e. g. the computer used for accessing the application) also does not need to be a particularly modern high-end machine. For further developments and especially for hosting more projects online, a dedicated server would be bought to ensure high computing power and provide an accurate service.

  • 8 PHP is an easy to learn and very popular script language that is available on almost every web host (...)
  • 9 MySQL is a dialect of SQL (structured query language), which is a language for creating databases a (...)
  • 10 PHP and MySQL are delivered with several standard setup files. They contain many different settings (...)
  • 11 Unicode is a character set that currently supports more than 110,000 characters and signs. In theor (...)

35The application’s two main services are executed by Apache and its integrated standard libraries: PHP8 and MySQL9. The setup of both is set to “medium”,10 except the maximum amount of memory, which is set to 512 MB each. All connections are encoded in Unicode11 (utf-8) by default; the data in MySQL is stored as utf-8 binary in order to support case-sensitive searches. Except for the images, all data is stored in one MySQL database.

36PHP serves as interface between the user’s actions and the database.

  • 12 HTML (Hypertext Markup Language) is the basic language that organizes and displays the contents of (...)
  • 13 JavaScript is a (usually browser-inside) script language that was created to extend HTML (http://ww (...)

37The front-end (e. g. the user-interface) is realized in HTML512 that is mostly generated and modified by JavaScript.13 The web pages are all written in PHP and mainly contain the output of the basic HTML-code and some JavaScript statements, as well as SQL-statements and file system operations for uploading, moving, and generating images.

38Since all major modern browsers support HTML and JavaScript, and since ImAnTo does not need any other plug-in to operate, there is no need for these or other software on the client. The client does not need to be a high-end computer since HTML and JavaScript are processed quite fast (on average) by computers that are no older than five years. That is one main advantage over other image annotations software that need plug-ins and/or a higher computational power. Another advantage is that also the languages used on the server are open source, well documented, and supported by a huge community.

39The application as result of the current project phase will be published with the techniques given above.

  • 14 Perl is a script language created by the American linguist Larry Wall. It is often used for linguis (...)
  • 15 A regular expression (abbr. regex and regexp) is a formal expression that is mainly used to check t (...)
  • 16 PostgreSQL is another dialect of SQL (http://www.postgresql.org). It supports regular expressions m (...)

40For further development, the script language would be changed to Perl,14 due to its better support and performance with regard to regular expressions.15 Additionally, PostgreSQL16 would replace MySQL for the same reason. These two modifications would not take a lot of time, since the languages are very similarly. Perl and PostgreSQL are open source, too.

5. FUTURE DIRECTIONS

41The parts presented so far will be completed by December 2013. This includes a user-friendly layout and menu entries for adding/deleting fields, lines, etc. The parts described in 5.1, 5.2, and 5.3 will be written after the current first project cycle is completed in December 2013, i. e. if additional funding is granted. These features will probably take another six months of development, so that they would be available for use in Summer 2014. The others described beginning with 5.4 cannot be handled within this timeframe — because of their complexity, these exceed the requirements of the additional funding. These are merely possible further developments, which would make ImAnTo the first and only annotation software able to annotate texts, images, videos, and sounds online by using a common browser.

5.1. USER RIGHTS MANAGEMENT

42A user rights management system would allow project leaders to restrict access to the whole project or single parts of it to specific users, groups, or the general public. As such, a project could be split into several chunks, each of which could then be assigned for completion by a specific user or group that could even be spread over different cities or countries. All other users would be locked out. For demonstration purposes and/or in order to avoid accidently deleting vital data, one could be allowed to login with limited user rights to view the content, but not to change it.

43For registering and accessing ImAnTo, one would only need to enter a valid e-mail address and a password. Each user would be identified via his name and e-mail address, so that e. g. the project leader (s) can put him into groups, etc.

5.2. LEXICON/DICTIONARY

  • 17 Used abbreviations: DEM = demonstrative; PL = plural; f = feminine

44Because of the freedom it offers for annotating various forms of handwritten materials, ImAnTo can store any textual information that is needed in arbitrary fields to an image or parts of it. For example, for the manuscript showing the modern Portuguese phrase estas lindas aves brancas, ImAnTo would contain the following annotation fields:17

45By the help of a built-in lexicon, researches would need to fill in the information of the lines glossing _ 1, glossing _ 2 and translation only once for their first occurrence. Afterwards, the lexicon would automatically retrieve the entered information for these words for all later occurrences. Accordingly, lindas in another line or manuscript would automatically be annotated as given above. This would not only save time in the process of annotation, it would also prevent spelling mistakes in different entries, e. g., accidently typing wihte instead of white for the next occurrence. However, note that if the very first entry contains an error, all others will of course also contain it. Such errors can easily be corrected by changing the lexicon’s entry and refreshing the annotations for updating the changes. That is, errors would need to be corrected only once.

46Additionally, the lexicon would also allow researchers to save all other necessary details associated with one word (e. g. POS, derivations, etymology, etc., as well as information about ambiguity). This information could be automatically added to the annotated page, either as sublines or as a separate info box linked to a particular field.

5.3. CITATIONS

  • 18 A URI is a unique identifier for resources of any kind, mainly used for web addresses or e-mail add (...)

47In publications, references are necessary in order to document the source of a given statement. Online editions of any kind are usually not designed for citation. Hence, there is the need for citation functionality that allows researchers to publish their annotations in papers. For doing so, the annotation would be copied to a special server hosting the quotable annotations via a Uniform Resource Identifier (URI).18 Such an annotation published for quotation cannot be edited, which guarantees that its content can be reliably used in papers. However, this also means that any errors will remain in the published annotations – just like errors will remain in editions of books or single papers once they are published. The only way of error-correcting would then be to create a new process of “publish for quotation” and use the resulting URI in papers from then on. Of course, if no publication ever contains the erroneous unique link, no one will ever have access to it, since it is not possible to browse through quotable annotations or guess/calculate their link by any algorithm. That is, for better or for worse, without the URI nobody is able to access the material published for quotation.

5.4. EXTENDED SEARCH FUNCTION

48Currently, the search function does not always find the suspected entries because of unclear word-boundaries. This is caused by search commands that do not yet consider special means for annotating abbreviations, e. g., the curly brackets in iē{rusa}l{em} isasa (Old Georgian “of Jerusalem”) surround text that is not written in the manuscript because the word “Jerusalem” was commonly abbreviated. The brackets are not considered characters in any coding system. Thus, they actually mark a word-boundary and, therefore, the search for iērusalem* will not find any results.

49To solve this problem, an extension to the current Search Page would need to be developed in which characters and punctuation marks could be defined by the user himself. Similarly, there would need to be a way to ignore non-characters in the middle of a word (i. e., defined as “several characters without spaces between them”). Only then would the search for iērusalem * also be able to find iē{rusa}l{em}isasa.

5.5. IMPORT/EXPORT

50Since the Caucasian palimpsests mentioned at the beginning of the paper are already transcribed and translated, their content will be imported by a separately written function. This function will not be a part of ImAnTo itself, but rather as an aid only for the specific project “Caucasian Palimpsests”. For all other projects, there should be an equivalent function, regarding the special formats researchers might use for their particular data. Thus, there is the need for an import function that allows users to import texts (and annotations) and the coordinates of associated shapes, both from different sources, e. g., XML, Toolbox, ELAN, plain text, etc. It is not possible to satisfy everybody’s needs concerning this matter, but many users would be able to easily upload their data into ImAnTo with a few clicks and extend it with additional information by hand. Without such a function that supports different formats and structures, researchers would have to manually convert the existing data to one specific XML scheme (since XML and XML-like structures are well-suited for this purpose). Manually converting the data would not be very user-friendly, however, so it would be ideal to implement such a function to do so automatically. Otherwise, most researchers of the Humanities would not be able to import their data because they might not necessarily have skills in software development and/or converting data structures. The same argument applies to exporting data.

5.6. OPEN STANDARDS FOR REPRESENTATION OF DATA

  • 19 SQL (Structured Query Language) is a programming language for creating databases, inserting data in (...)
  • 20 http://www.openannotation.org/

51No open standard for data representation is used at the moment. All information is organized in a self-written structure that is stored in a SQL-database.19 I plan to do more research into open standards, e. g. the Open Annotation Data Model,20 to check whether it can be used for representing all levels, lines, and relationships needed for typical palimpsests.

5.7. AUDIO/VIDEO ANNOTATION

52In general, all kinds of data can be annotated by computer programs. Even webapplications like the one presented are able to do so, and they are becoming better with time because of the more powerful features afforded by computer languages and their modifications.

53As mentioned in chapter 4, HTML5 is the upcoming standard of the most-used language for creating websites. Thanks to the fact that it supports audio and video files inherently, no plug-ins for playing those files would be needed. Consequently, materials that contain audio comments, speeches, any content of field research, the documentation of excavations or church paintings, etc., could be split into sections/chapters and complemented with all kinds of annotations. During play back, ImAnTo would automatically change the annotations in relation to the file’s current position in the timeline.

54Expanding the development in this direction would result in the first and only online universal multimedia annotation software. Though this further functionality is planned, it is not clear whether it will be possible to realize in the future.

6. CONCLUSION

55From the first month of preparation up until now, almost every step of development has been implemented without problem, as described in chapter three. There still are a few bugs present, but I am convinced that they will be fixed soon so that the development will be successfully finished by December 2013.

56Nevertheless, the bugs do not prevent the use of ImAnTo in its current state: researchers can already draw shapes on images, enter annotation data in as many lines and levels as needed, save all this information, and search through it. The importation of already existing data for the Caucasian palimpsests (i. e. transliterations, translations, and comments) is ready, too. Furthermore, ImAnTo has been chosen for use in a separate project focusing on Old Lithuanian manuscripts. In addition, it is currently being considered for use as the annotation tool in an archaeology project at Goethe-University Frankfurt that is currently in its planning phase.

57On the one hand, because of ImAnTo’s simplicity, users will find it easy to learn to use. Furthermore, the modern and powerful techniques used in the application, its powerful functions, its optimization for palimpsests, and its low-level overhead on the other hand, clearly make it very attractive as the tool of choice for annotating palimpsests, manuscripts, paintings, coins, etc.

58In a nutshell, the future appears positive and promising for the continuing development and use of ImAnTo.

Bibliographie

BIBLIOGRAPHY

Since this paper is documenting the development of a computer program, there was no need for citing printed or digital sources.

The web-resources given in the technical explanations where accessed last on 21st of June 2013.

Notes

1 http://www.proloewe.de/en/home/welcome.html
http://verwaltung.hessen.de/irj/HMWK_Internet?uid=fa560c0b-ed11-9311-1010-43bf5aa60dfa
http://www.digital-humanities-hessen.de/
http://www.digital-humanities-hessen.de/iii-archiv-corpus-edition/35-handschriftenanalyse-undedition/49-kaukasische-palimpseste

2 Unicode is a standard for encoding and displaying characters and symbols in most of the world’s languages. It currently supports more than 110,000 characters/symbols. See also: http://www.unicode.org.

3 Modern hard disks in workstation computers (e. g. office computers) usually read around at least 60 Megabytes per second (MB/s), i. e. tens of thousands of pages of full text. Servers, however, are able to read with a transfer rate of even 600 MB/s or more (http://en.wikipedia.org/wiki/RAID; http://en.wikipedia.org/wiki/Serial_ATA).
Modern central processing units (CPUs) are able to execute some billion “thoughts” (http://en.wikipedia.org/wiki/Instructions_per_second). Considering this, a human being can never compete with computers when it comes to analyzing text.

4 XML (Extensible Markup Language) is a markup language that is used mainly for hierarchically saving information readable to humans and computers alike. Each XML-document consists of a tree-structure. Due to the fact that there are only a few limitations in such a structure, one is able to represent the data almost unrestrictedly. Most programming languages can access XML-documents via freely available libraries. Despite its structuring, XML is mainly used for exchanging information from one system or structure to another (http://www.w3.org/standards/xml).

5 I recommend Mozilla Firefox (http://www.mozilla.org) for several reasons including modernness, security and speed; however, Opera, Chrome, Safari or MS Internet Explorer are also acceptable alternatives that work with the application.

6 Open source software is free of charge and anyone is allowed to create his own version by modifying the (open) source code. This is a security advantage since anyone and everyone can check the code for harmful commands. For this reason, security gaps in open source programs generally do not remain undiscovered for very long.

7 Apache is a very popular and powerful free web server that enables one to involve database languages, programming languages, security techniques and self-created programs for websites and web-applications (http://apache.org). It currently runs over 300 million websites all over the world (http://news.netcraft.com/archives/2013/03/01/march-2013-web-server-survey.html).

8 PHP is an easy to learn and very popular script language that is available on almost every web hosting service (http://php.net). It can be used for generating and submitting HTML and JavaScript-code, accessing databases and the file system.

9 MySQL is a dialect of SQL (structured query language), which is a language for creating databases and their entries, as well as for accessing them. It is very common since it is free, easy to learn and administrate, and very powerful (http://dev.mysql.com).

10 PHP and MySQL are delivered with several standard setup files. They contain many different settings in order to fit to the system’s hardware.

11 Unicode is a character set that currently supports more than 110,000 characters and signs. In theory, it is able to display more than 1 million characters and signs in its current specification. It is the set of choice in linguistics since it ensures that all researchers can access the data, regardless of their local settings.

12 HTML (Hypertext Markup Language) is the basic language that organizes and displays the contents of websites. Its current standard is version 4.01 (http://www.w3.org/standards/webdesign/htmlcss). The upcoming version, HTML5, is currently under development and only partly supported by modern major browsers (http://www.w3.org/TR/html5). It is a very powerful new version of this language, supporting several multimedia files and actions that currently are not usable in pure websites (i. e. without plug-ins). Nevertheless, the most important feature for ImAnTo is already supported by all browsers, i. e., displaying the element canvas for working with images and drawing shapes on them. HTML5 will become a standard and all browsers will support the new elements and features some short time after the release of the standard.

13 JavaScript is a (usually browser-inside) script language that was created to extend HTML (http://www.w3.org/standards/webdesign/script). By aid of JavaScript it is possible to modify existing HTML elements on a website dynamically, as well as to add them.

14 Perl is a script language created by the American linguist Larry Wall. It is often used for linguisticrelated programming and very powerful in using regular expressions, since they are an inbuilt part of the language (http://www.perl.org).

15 A regular expression (abbr. regex and regexp) is a formal expression that is mainly used to check text patterns. By means of its many options and supported character classes one can quite easily search even thousands of pages of text for complex patterns within seconds (http://en.wikipedia.org/wiki/Regular_expression).

16 PostgreSQL is another dialect of SQL (http://www.postgresql.org). It supports regular expressions more efficiently than MySQL. However, it is more difficult to administrate.

17 Used abbreviations: DEM = demonstrative; PL = plural; f = feminine

18 A URI is a unique identifier for resources of any kind, mainly used for web addresses or e-mail addresses (http://en.wikipedia.org/wiki/Uniform_resource_identifier).

19 SQL (Structured Query Language) is a programming language for creating databases, inserting data into them, and requesting these data (http://en.wikipedia.org/wiki/SQL).

20 http://www.openannotation.org/

Table des illustrations

Titre Figure 1 – Thumbnails of all images of the selected project
URL http://books.openedition.org/cidehus/docannexe/image/1093/img-1.jpg
Fichier image/jpeg, 528k
Titre Figure 2 – Preview of the selected image
URL http://books.openedition.org/cidehus/docannexe/image/1093/img-2.jpg
Fichier image/jpeg, 180k
Titre Figure 3 – Screenshot showing part of the Annotation Page
URL http://books.openedition.org/cidehus/docannexe/image/1093/img-3.jpg
Fichier image/jpeg, 344k
Titre Figure 4 – Screenshot showing part of an annotation; focused on character-based first sublevel
URL http://books.openedition.org/cidehus/docannexe/image/1093/img-4.jpg
Fichier image/jpeg, 44k
Titre Figure 5 – Search Page with highlighted results
URL http://books.openedition.org/cidehus/docannexe/image/1093/img-5.jpg
Fichier image/jpeg, 216k
URL http://books.openedition.org/cidehus/docannexe/image/1093/img-6.jpg
Fichier image/jpeg, 47k

Auteur

Institute of Empirical Linguistics & LOEWE Center “Digital Humanities”
Goethe-University Frankfurt am Main
Investigador do Loewe Institut – Center Digital Humanities, Frankfurt am Main, Alemanha. Especialidades: Linguística Empírica, produção de ferramentas para a criação de corpora no âmbito de projectos de Humanidades Digitais. raaf@em.uni-frankfurt.de

© Publicações do Cidehus, 2013

Conditions d’utilisation : http://www.openedition.org/6540

Lire

Open access

Acheter

Volume papier

Chargement

Unavailable