Work Flows and Wish Lists: Reflections on Juxta as an Editorial Tool

I have had the opportunity to use Juxta Commons for several editorial projects, and while taking a breath between a Juxta-intensive term project last semester and my Juxta-intensive MA thesis this semester, I would like to offer a few thoughts on Juxta as an editorial tool.

For my term project for Jerome McGann’s American Historiography class last semester, I conducted a collation of Martin R. Delany’s novel, Blake, or, The Huts of America, one of the earliest African American novels published in the United States.Little did I know that my exploration would conduct me into an adventure as much technological as textual, but when Professor McGann recommended I use Juxta for conducting the collation and displaying the results, that is exactly what happened. I input my texts into Juxta Commons, collated them, and produced HTML texts of the individual chapters, each with an apparatus of textual variants, using Juxta’s Edition Starter. I linked these HTML files together into an easily navigable website to present the results to Professor McGann. I’ll be posting on the intriguing results themselves next week, but in the meantime, they can also be viewed on the website I constructed, hosted by GitHub: Blake Project home.

Juxta helped me enormously in this project. First, it was incredibly useful in helping me clean up my texts. My collation involved an 1859 serialization of the novel, and another serialization in 1861-62. The first, I was able to digitize using OCR; the second, I had to transcribe myself. Anyone who has done OCR work knows that every minute of scanning leads to (in my case) an average of five or ten minutes of cleaning up OCR errors. I also had my own transcription errors to catch and correct. By checking Juxta’s highlighted variants, I was able to—relatively quickly—fix the errors and produce reliable texts. Secondly, once collated, I had the results stored in Juxta Commons; I did not have to write down in a collation chart every variant to avoid losing that information, as I would if I were machine- or sight-collating. Juxta’s heat-map display allows the editor to see variants in-line, as well, which saves an immense amount of time when it comes to analyzing results: you do not have to reference page and line numbers to see the context of the variants. Lastly, Juxta enabled me to organize a large amount of text in individual collation sets—one for each chapter. I was able to jump between chapters and view their variants easily.

As helpful as Juxta was, however, I caution all those new to digital collation that no tool can perfectly collate or create an apparatus from an imperfect text. In this respect, there is still no replacement for human discretion—which is, ultimately, a good thing. For instance, while the Juxta user can turn off punctuation variants in the display, if the user does want punctuation and the punctuation is not spaced exactly the same in both witnesses, the program highlights this anomalous spacing. Thus, when 59 reads

‘ Henry, wat…

and 61 reads

‘Henry, wat…

Juxta will show that punctuation spacing as a variant, while the human editor knows it is the result of typesetting idiosyncrasies rather than a meaningful variant. Such variants can carry over into the Juxta Edition Builder, as well, resulting in meaningless apparatus entries. For these reasons, you must make your texts perfect to get a perfect Juxta heat map and especially before using Edition Starter; otherwise, you’ll need to fix the spacing in Juxta and output another apparatus, or edit the text or HTML files to remove undesirable entries.

Spacing issues can also result in disjointed apparatus entries, as occurred in my apparatus for Chapter XI in the case of the contraction needn’t. Notice how because of the spacing in needn t and need nt, Juxta recognized the two parts of the contraction as two separate variants (lines 130 and 131):

spacing 2

This one variant was broken into two apparatus entries because Juxta recognized it as two words. There is really no way of rectifying this problem except by checking and editing the text and HTML apparatuses after the fact.

I mean simply to caution scholars going into this sort of work so that they can better estimate the time required for digital collation. This being my first major digital collation project, I averaged about two hours per chapter (chapters ranging between 1000 and 4000 words each) to transcribe the 61-62 text and then collate both witnesses in Juxta. I then needed an extra one or two hours per chapter to correct OCR and transcription errors.

While it did take me time to clean up the digital texts so that Juxta could do its job most efficiently, in the end, Juxta certainly saved me time—time I would have spent keeping collation records, constructing an apparatus, and creating the HTML files (as I wanted to do a digital presentation). I would be remiss, however, if I did not recommend a few improvements and future directions.

As useful as Juxta is, it nevertheless has limitations. One difficulty I had while cleaning my texts was that I could not correct them while viewing the collation sets; I had, rather, to open the witnesses in separate windows.

screenshot windows

The ability to edit the witnesses in the collation set directly would make correction of digitization errors much easier. This is not a serious impediment, though, and is easily dealt with in the manner I mentioned. The Juxta download does allow this in a limited capacity: the user can open a witness in the “Source” field below the collation visualization, then click “Edit” to enable editing in that screen. However, while the editing capability is turned on for the “Source,” you cannot scroll in the visualization—and so navigate to the next error which may need to be corrected.

A more important limitation is the fact that the Edition Starter does not allow for the creation of eclectic texts, texts constructed with readings from multiple witnesses; rather, the user can only select one witness as the “base text,” and all readings in the edition are from that base text.

screenshot Edition Starter

Most scholarly editors, however, likely will need to adopt readings from different witnesses at some point in the preparation of their editions. Juxta’s developers need to mastermind a way of selecting which reading to adopt per variant; selected readings would then be adopted in the text in Edition Starter. For the sake of visualizing, I did some screenshot melding in Paint of what this function might look like:


Currently, an editor wishing to use the Edition Starter to construct an edition would need to select either the copy-text or the text with the most adopted readings for the base text. The editor would then need to adopt readings from other witnesses by editing the the output DOCX or HTML files. I do not know the intricacies of the code which runs Juxta. I looked at it on GitHub, but, alas! my very elementary coding knowledge was completely inadequate to the task. I intend to delve more as my expertise improves, and in the meantime, I encourage all the truly code-savvy scholars out there to look at the code and consider this problem. In my opinion, this is the one hurdle which, once overcome, would make Juxta the optimal choice as an edition-preparation tool—not just a collation tool. Another feature which would be fantastic to include eventually would be a way of digitally categorizing variants: accidental versus substantive; printer errors, editor corrections, or author revisions; etc. Then, an option to adopt all substantives from text A, for instance, would—perhaps—leave nothing to be desired by the digitally inclined textual editor. I am excited about Juxta. I am amazed by what it can do and exhilarated by what it may yet be capable of, and taking its limitations with its vast benefits, I will continue to use it for all future editorial projects.

Stephanie Kingsley is a second-year English MA student specializing in 19th-century American literature, textual studies, and digital humanities. She is one of this year’s Praxis Fellows [see Praxis blogs] and Rare Book School Fellows. For more information, visit, and remember to watch for Ms. Kingsley’s post next week on the results of her collation of Delany’s Blake.

Re-visualizing in Juxta:

The Juxta R&D team recently did a demo of Juxta Commons at a DH Day conference at NC State University, and one of the attendees brought the site to our attention. It’s a great resource, tracking changes to online, “published” articles from some of the largest media outlets out there. But, while NewsDiffs brings together a bunch of different versions of these online articles, their visualizations are only helpful for two versions at a time.

As an experiment, I took several versions from this example and plugged them into Juxta Commons. When combined in the heat map, the results were truly surprising.

In this example, the newest version (captured on November 6, 2012) is the base witness, with the previous revisions made on the day of the articles release included in the set. Just imagine: readers visiting the New York Times article at 11:15am would have read a very different set of opening paragraphs than those checking in at 11:45am.



A Preview of Juxta Commons

The NINES R&D team is happy to announce a new phase of testing for Juxta online: Juxta Commons. We’re entering our final phase of intensive testing on this new site for using Juxta on the web, which breaks down the processes of the desktop application so you always have access to your raw source files and your witnesses, in addition to your comparison sets. We’ve even added more ways to work with XML, as well as an option to import and export files encoded in TEI Parallel Segmentation.

We have invited a group of scholars and users to try out Juxta Commons for the next two months, and share their experiences online. They’ll be exploring new ways to add source files, filter XML content and browse our newly-updated visualizations, previewed in the gallery above.

If you would like to be a part of this group of testers, please leave a comment below, and we’ll get in touch with you.

Beta-release of Juxta includes online sharing

Calling all beta testers!

Over the past few months, NINES and the developers of Juxta have been busy adapting the application for use on the web. In order to expand our testing capabilities, we’re releasing a version of the desktop client that offers users the ability to share comparison sets online.

If you have any sets of witnesses to a particular work that you would like to collate and share, we invite you to sign up and download the beta version  to try out some of our online features. Please keep in mind that this is a trial version of the web-service, and may be subject to changes and updates over the next few months. Joining us now ensures that your feedback will make the full release of the software better than we could manage in-house.

 Please help us make Juxta better!

Working with non-Roman alphabets in Juxta

Now that Juxta 1.3 has been refined and released, the development team at NINES has been discussing new directions for the software. First and foremost is the adaptation of Juxta’s collating power for texts in languages other than English. Comparisons of texts in French and Italian work pretty well, but we’re still investigating the necessary diacritics to make such operations more exact. However, it seems that scholars working with non-Roman alphabets have been left out of the conversation.

Do any Juxta users out there have any experiences with foreign language collation to share with us?

server hiccups

For all those who had trouble accessing the site this week, we’re happy to announce it’s up and running again! We apologize for the delay and encourage you to access the manual and software download pages once more.

Juxta 1.3 Released

Juxta 1.3 is now available for download here. It has the following new features:

1) Search over all documents.

Juxta Search
A search box has been added to the toolbar, making it possible to find instances of a word or phrase within all documents in the comparison set. Those results are listed in the Search pane at the bottom of the screen (see image above). Clicking on a line in that pane will display the document and the search results. Note that Juxta will remember the last searches that were performed and show them in the search drop down list.

2) Line numbers appear for the witness and base texts.

Line Numbers

Now, when the “toggle line numbers” menu item is selected, the line numbers appear alongside the witness text, in addition to the those coresponding to the base text.

3) “Moves”: the ability to correlate similar passages that are differently located in two documents.

Juxta Moves

The Passages feature from the last version has been reworked into the new, “Moves” feature. In the side-by-side collation view, the user may select text in both the base and the witness documents representing a passage identified as having moved (1). The move button (2) will become enabled at that point.

Juxta Move Completed
Click here to create the move. You will see an outline of the passages (3) and a line connecting them, with an entry made in the Moves pane (4). Clicking the entry brings the move into view.

Altogether, these features represent a significant improvement to Juxta as a textual collation tool. Download it and give it a try today!

Juxta 1.2.2 Released

Juxta 1.2.2 is now available for download. The major new feature in this release is an improved fragment selection mechanism and the ability to easily preview files before collating them. This functionality is accessed via the “Files” tab on the left hand panel, depicted below.

Clicking on the “Files” tab brings up a tree of the files in the currently selected base directory. Clicking on the file icon allows the scholar to select a directory from which to select files for collation.

Double clicking on files with a “txt” or “xml” extension opens them in a preview mode. The scholar can then choose to import the entire file into the collation or to highlight a fragment and pull just the highlighted fragment into the collation. Fragments carry with them the metadata and lineation from the source text, if any. This new functionality replaces the old fragment selection mode with a more integrated solution.

Juxta 1.2.1 Released

It has been a while since we have had an update on this blog, but work has been occurring this year behind the scenes at ARP. Juxta 1.2.1 is now available for download.

Last summer, Performant Software Solutions developed a new version of Juxta for ARP. This is the version I demoed at COST 32 Workshop in Antwerp, Belgium in September. Below is a summary of the new features and bug fixes found in this release.

New Features

Passage Collation – Juxta can now collate texts in which passages appear in different order from one text to the next. A new user interface component, the Passage Panel, guides the user through the process. See the updated user’s manual for more information.

Fragment Collation – Juxta can also collate fragments of texts. This is useful when the target of collation is embedded in a document.

Free Scrolling in Comparison View – Documents in the side-by-side comparison view can now be scrolled independent of one another.

Easier to Find Samples Directory – A file menu shortcut has been added that takes you straight to the samples directory.

Improvements to generation of Critical Apparatus – A dialog box now allows you to specify a title for the critical apparatus. A progress bar is now provided while the critical apparatus is being generated. Some bugs with the generation of lemmas within the apparatus have been fixed.

Find Works in Either Document – The Find Dialog can now find text in either document in the side-by-side comparison mode.

Bug Fixes

Improvements to Collation Algorithm – Hans Walter Gabler found some problems with the collation output in specific circumstances. We have corrected these errors.

Margin Box Clipping – Occasionally, when the window was resized it would cause the margin boxes to be clipped, this has been fixed.

Scrollbar Positioning Sometimes Incorrect – When loading a large document the scrollbar would become positioned incorrectly, this has been fixed.

Image Display Not Updating – When flipping between documents, sometimes the image associated with the document on screen would not appear, this has been fixed.

Large Images Cause Scrolling to Be Erratic – Related to the issue above, scrolling through documents which have images associated with them could be jerky at times as the images loaded. Loading has been adjusted so this no longer occurs.

Out of Memory Error – Loading two large, completely unrelated documents could cause a system error. This has been fixed.