Have MLIS, Will Travel – Quick note

I’ve been slowly working out the kinks on my new job-sharing site, Have MLIS, Will Travel. While it is not quite ready for general use right now, it is looking very promising

In my test phase, since the API calls to the Google Places API contain so much data, I’ve been saving and logging the Latitude/Longitude of the city for each job posting. I’m attaching a quick visualization for a small selection of the data collected over the last month.

https://batchgeo.com/map/d4dbd9868697acca508ca651a0dbdc54

 

Advertisements

The BookBox : A localized, portable ebook delivery system using open-source components

Ebooks are becoming an essential part of library services. Most library eBooks are delivered through third party vendors encumbered by DRM and access controls. Open and out of copyright ebooks provide a large library of books and articles that Librarians and information professionals can experiment with and provide to their patrons.

This post shows how librarians can use cast-off technology to create a localized ebook  delivery system to deliver ebooks over a Wifi connection. The system described in this post can be put together with a wide variety of hardware, much of which can be re-purposed from obsolete components. This system can also be powered by battery power to deliver a large number of electronic texts to school rooms, agricultural centers or higher educational institutions even in areas without consistent electrical power.

Components of the system.

  1.  A computer that can run the Debian operating system. Old laptops, old desktops, plug computers and even NAS devices like the PogoPlug can be used as the computing platform.
  2.  A Router capable of running  the DD-WRT open source firmware
  3.  A USB drive or flash memory key ( if computer doesn’t have memory like the PogoPlug)
  4.  (optional) A battery pack
  5.  (optional) An enclosure to hold the system.

For my example build I used a Seagate FreeAgent DockStar as the base computer and a Asus WL – 330GE as the router. My example build fits in a standard metal lunch box, including a battery pack. My example build cost ~$100 US.

Plug the computer into a router with an ethernet cable.

Install Debian on the computer platform. Debian can be downloaded from http://www.debian.org/.

In my example build, I used a Seagate DockStar FreeAgent NAS device and installed Debian from the version maintained by Jeff Doozan. His instructions for installing Debian on the Dockstar are found at http://jeff.doozan.com/debian/

If using the DockStar as I did in my example build, David Dart’s guide to creating the ‘PirateBox’ can give you additional valuable advice on mounting partitions on your drive or USB memory key.

http://wiki.daviddarts.com/PirateBox

Install the Monkey webserver software on your project computer. http://monkey-project.com/
The Monkey webserver is an extremely simple webserver with minimal configuration required. You can install monkey using the apt program manager by issuing the command

sudo apt-get install monkey

All of the monkey configuration files are found in conf/ directory.
Open monkey.conf in nano by issuing the command

sudo nano monkey.conf

In monkey.conf, ensure that monkey is listening on port 80, and note the location of the HTTP directory. If necessary, you can change the location of the HTTP directory. This is the location where we’ll be uploading the ebook files.

Save the monkey.conf file.

To test, create an index.html file with ‘Hello World’. Open the IP address that your router has given the computer like this http://%5BIP Address]:[Port Number].

On another computer, use the Calibre software (http://calibre-ebook.com/ ) to create an ebook library from the source of your choice. Calibre is a powerful tool for organizing, editing, creating and distributing eBooks. Sources for your library include the Project Gutenberg books, open access books from the Internet Archive, or a project like TEEAL http://www.teeal.org/. Using Calibre you can add metadata, download covers and other details from Amazon.com and the Internet Archive.

Once your Calibre library is created, you can use a companion program called Calibre2odps to create a set of static HTML pages. http://wiki.mobileread.com/wiki/Calibre2opds

This set of static HTML pages can be uploaded to any webserver, in any location.

Copy the complete Calibre2odps library over to the HTTP directory of your Debian webserver.

Your webserver can now be plugged into any network connection, and will serve up the complete catalog in HTML and ODPS formats on port 80.

I created a simple index.html landing page pointing to both the catalog.html and catalog.xml files. This is the default landing page.

To create a standalone system, use an old router capable of running the DDWRT open source firmware. You can find a list of compatible routers by visiting http://www.dd-wrt.com/site/index.

Install DDWRT according to the directions for your particular model. For my project I used the Asus WL330GE because it can be powered by a USB cable.

Once DDWRT is installed, you can change the SSID to something more appropriate, I suggest ‘Bookbox’

Plug the DDWRT router into your computer platform, turn everything on and connect to your ‘Bookbox’ wireless network.

The goal of this setup is to force all traffic to visit index.html when they connect to the network.

Point your browser to 192.168.1.1 to visit the control interface for the router.  Go to Status > LAN and write down the IP address that has been assigned to the computer by the router. This may take a while.

Next, go to Services > Services  and set the DNSMasq and Local DNS buttons to Enable. In the “Additional DNSMasq Options” add the following

address=/#/0.0.0.0 (enter the IP address you wrote down)

Click Save.

Go to Services > Hotspot tab and click Enable beside HTTP Redirect.

Enter the following items

HTTP Destination IP: 0.0.0.0 (Your server’s IP Address)

HTTP Destination Port: 80

HTTP Source Network: 192.168.1.0 (the “0” denotes the whole range of IP addresses under this address)

Click Save, then go to the Administration > Management tab and click the Reboot Router button at the bottom of the page. Wait for the router to reboot and enter any address in the address bar of your browser. You should now be redirected to the index.html landing page .

If you connect to the Bookbox wireless network with a ODPS compatible ebook reader you can browse a catalog of ebooks within the ereader.

If this system has been put together with a plug computer, or using one of the systems that David Darts prototyped with his Piratebox systems, you should be able to power the entire system using a consumer laptop external battery pack. The DockStar or Pogoplug systems are particularly easy to power as their extra USB ports can easily provide power to the router over USB. The entire system, when set up using the DockStar and the Asus WL330GE  can be fit into a standard tin lunchbox, including a fairly large battery pack.

I have found that simpler web browsers, such as those found on e-readers actually handle this system better than the more complex Mobile Safari. From the feedback from David Darts project, it looks like Android phones seem to handle this connection quite well.

Resources

http://wiki.daviddarts.com/PirateBox_DIY_Plug_Computer
http://jeff.doozan.com/debian/
http://calibre-ebook.com/
http://wiki.mobileread.com/wiki/Calibre2opds
http://monkey-project.com/documentation/index

The Architecture of the Next Gen ILS

Marshall Breeding’s Systems Librarian column in the October 2010 edition of Computers in Libraries has been on my mind extensively recently. I’ve spent the last two years integrating our ILS with a hosted Aquabrowser instance and Summon and I was pretty surprised at the amount of ‘hacking’ that had to go on to make things work.

The primary method we are using in the Library world is MARC 21 export for these services. Why don’t these services support our existing search infrastructure of Z39.5?

Is this a speed issue? A caching issue? It just seems quite silly to depend on a frail script to upload large files on a regular basis when the infrastructure to support queries already exists and has been the standard for years.

Barcodes are dead. Long live the barcode!

I’ve noticed a mindless consensus among the geek literati when talking about QR Codes. Often a blog post will discuss the various uses they are put to, then end with the phrase ‘But I think that NFC will replace them in the near future’.

Bullshit.

I have an RFID system in my library. My long term experience in libraries is with barcodes. I’ve already done the QR Code vs NFC dance. While both have their advantages, I think the two technologies are complementary, not competitive. And barcodes have the advantage.

The cost of production for barcodes is ink and paper. They have a standard protocol for communication between the reader and the computer. I can plug a barcode reader into a USB port and scan off the raw numbers on the barcodes into Excel.

I can make a barcode whenever I want using standard office products. If I’m enthusiastic I can decode what a barcode means with a Mark 1 Eyeball. This simplicity of communication between printed text and computer systems is the major advantage of barcodes

The major problems with barcodes is the bandwidth and the data storage. Laser scanning is a little slow, and the few bits you can store in a printed barcode are not really sufficient.

Using barcodes as identifiers, then pulling the rest of the information from a networked system solves the bandwidth and information storage problem.

RFID tags cannot just replicate what barcodes do. If I’m using a NFC tag to point to a URL, what’s the point? The $1 NFC tag sticker pointing to that URL is redundant when it can be duplicated by my printer as part of the standard design for a few cents in design fees and ink.

RFID will work in a few crucial areas.

  1. The offline data requirements are of a larger size than what can be accommodated in a barcode.
  2. You need to avoid the one-at-a-time scan bottlenecks of barcodes.
  3. You need to track the location of a tagged item at all times.
  4. You have two networked devices (say POS and smartphone) that need to communicate securely only in a specific place. (and Facecash proves that QR codes work just fine for this)

RFID stickers will not replace the barcode.

NFC Point of Sale devices have promise, and are actually in use already with many tap-to-pay services. Industry and government will continue to use RFID in the locations where it works. NFC will probably revolutionize device-to-device communication.

For you and me, the easiest way to get information from print to computer will still be some variation of the barcode.

Related Editions and Associated Information in Integrated Library Systems

This is based on a request from Tejaswi Tenneti in relation to an answer I gave a while ago here: Brett Williams’s answer to Ontologies: Is there an Ontology that models Enterprise products and associated knowledge articles?

This post will walk you through some of the advanced information processing that goes on in large library systems.

The basis of library records is a standard interchange format called the MARC record. The MARC record is an ISO standard that was originally developed by Henriette Avram at the Library of Congress in the 1960’s.

MARC is an interchange format, it’s a way to transfer information both offline and online, and was originally developed for tape-based processing of information, long before ARPA-net was a twinkle in a CRT screen.

There are several standard MARC formats, as well as several different implementations of MARC, both ASCII and UTF-8 encoded. For simplicity, I’m just going to show you what a bibliographic record looks like. A bibliographic record is filled in with the information about a single item.

Here’s the first couple of lines of a MARC record I’ve cracked open with a tool to show you the structure of the record.

=LDR 01437nam 2200373 a 4500
=001 9781849729390
=003 Credo
=005 20110223080808.0
=006 m\\\\\\\\d\\\\\\\\
=007 cr\cn|||||||||
=008 110211r20032011enka\\\\od\\eng\d
=020 \\$a9781849729390 (online)
=020 \\$z9780007165414 (print)
=020 \\$z0007165412 (print)
=035 \\$a(OCoLC)703091350
=035 \\$a(CaBNvSL)slc00226357
=035 \\$a(Credo)hcdquot2003
=040 \\$aCaBNvSL$cCaBNvSL$dCaBNvSL
=050 \4$aPN6081$b.D495 2003eb
=082 04$a080$222
=245 00$aCollins dictionary of quotations$h[electronic resource].

The raw record looks like this. Quora can’t even display some of the critical separation markers used. Like I mentioned, this was developed in the 1960s.

01437nam 2200373 a 45000010014000000030006000140050017000200060019000370070015000560080041000710200027001120200026001390200023001650350021001880350026002090350023002350400030002580500025002880820012003132450061003252500025003862600080004113000063004915060060005545200060006145300037006745380036007115880054007476500039008016550022008407100027008627760046008898560128009359781849729390Credo20110223080808.0m d cr cn|||||||||110211r20032011enka od 000 0 eng d a9781849729390 (online) z9780007165414 (print) z0007165412 (print) a(OCoLC)703091350 a(CaBNvSL)slc00226357 a(Credo)hcdquot2003 aCaBNvSLcCaBNvSLdCaBNvSL 4aPN6081b.D495 2003eb04a08022200aCollins dictionary of quotationsh[electronic resource].

And yes, before you ask, there is an XML version. It’s still catching on, but it’s looking like the future interchange format.

So, here’s my workflow. I get a new book in the library. I take a look at it, and then I search for the MARC record in another library, the Library of Congress or from a company called OCLC that has tools that make this process very quick. I download the record to my computer and open it up using tools in my ILS (integrated Library System). I check over the MARC record. It doesn’t have to be exact, but I do check the ISBN, the title, author and a few other points. I also have tools that check the author (is it Chester M Arthur or Chester M. Arthur?) title and some other fields to ensure consistency.

Then, I tell my ILS to add it to the catalog. At this point, the MARC record is broken up and put into tables in a database. My current ILS runs on Oracle, others use MySQL.

When the record is put into the database, it’s checked against the current holdings. If I already have this exact edition of the Collins Dictionary of Quotations, it will show up as another item, identical to the first. If I have an older edition, my new record will show up as another item, with a different ISBN and other details, and the ILS will show a unified record when I look up ‘Collins Dictonary of Quotations’ with both editions identified.

Is the item written by Mark Twain, who’s really Samuel Clemens? That decision got made years ago, so if you search the Harvard Library or the library in Minot, North Dakota you’ll get the same answer. Were there 15 different editions with different contents? There’s a rule for that.

Bringing this back to industry, say you are taking a product through it’s lifecyle. In R&D it’s Project Wiffleball. When it moves over to Product Development, it becomes Project Unicorn E456. When you release it to the public it’s called the E-Nebulizer, which you release versions E999, E678, E469, ET689 and SV368. What a library system could do is collate the information from

Project Wiffleball
Project Unicorn E456
E-Nebulizer E999
E-Nebulizer E678
E-Nebulizer E469
E-Nebulizer ET689
E-Nebulizer SV368

and show it all in the same place, on a single screen, with no duplication, and with clear links to knowledge objects that relate generally to the project and others that relate to specific versions.

Now, this is all done within the ILS. I can also pull information from outside sources to enrich the contents of the bibliographic record

Here’s a record from my catalog.

http://library.cna-qatar.edu.qa/?itemid=|marc-qatar|000060791

That section in the middle is the MARC record from my catalog. On the right is the book cover, drawn from a 3rd party service (Syndetics). It runs by grabbing the ISBN, running it through a collation service called xISBN to identify all of the editions. Then, if there is a cover for the given ISBN, it gives that cover. If Syndetics doesn’t have it, I get another edition’s cover.

Moving on down, the full table of contents, also drawn from Syndetics. This is indexed, along with all of the main record information and is searched. Further down, reviews of the book are also available, also keyword indexed and searchable.

All right! Now my library is very basic, let me show you something a little more complex.

Here’s the University of Southern Queensland Library in Toowoomba, about 3 hours from Brisbane. I had the pleasure of talking quite extensively with their IT team a couple of weeks ago at a conference.

Take a look at this record

http://library.usq.edu.au/Record/vtls000420789

You can see the Syndetics book covers over on the left hand side of the screen. Now, check out the Similar Items link.

Using a relevance algorithim, a list of similar works is pulled up and integrated with the individual record. This is a database maintained in-house by their IT department, it’s part of the library database that serves up this finding aid.

OK, let’s flip over to USQ’s mobile interface.
http://m.library.usq.edu.au/
Now try that top link again, you’ll have the right cookie now to view the mobile interface.

Check out the ‘Show on Map link’ This will load a graphic, sized for the iPhone that indicates the location of the book. This is from a database maintained by their IT department, separate from the library database that serves this finding aid.

OK, let’s hop back a little closer to home.

The San Diego Public Library has included QR codes in their book records using the Google Charts API

http://libpac.sdsu.edu/search~S0?/XTesting&SORT=D/XTesting&SORT=D&SUBKEY=Testing/1%2C13291%2C13291%2CB/frameset&FF=XTesting&SORT=D&1%2C1%2C

If you’ve made it this far, you’ve got a real interest in displaying information. As I mentioned before in my answer, this is a very rough introduction to a complex topic. If you’re really interested, and are looking for consultants, I have a few contacts in Knowledge Management, Libraries and the systems behind them that can give you much more help

Cross posted from Quora

eBook Distribution – Experiment 1

I’ve been experimenting with some different methods of distributing ebooks using existing infrastructure. These experiments are localized (thinking of the library as a place where people go) while at the same time using WiFi, networks, computers, web servers and other tools of the internet.

My first bare-bones experiment is using Mongoose

Mongoose is a radically simple webserver. Download the executable, put it in the folder you want to share, run it and you can access it at http://localhost:8080 or at http://%5BcomputerIP%5D:8080.

If you want to run it off of port 80, or add additional functions, you can edit the optional mongoose.conf file

With a small edit of the mongoose.conf file, enabling directory listing, you can serve up a folder over a network, allowing users to download whatever is inside that folder.

Here’s some advantages

1) This is a 2 minute project.You have a folder to share, and a way to share it.

2) This method respects your user’s privacy. With a quick setup like this, you’re not harvesting their usernames, reading preferences or creating advertising profiles.

3) You can set this system up anywhere, using almost any hardware from the last 10 years.

 

Disadvantages

1) It’s ugly. You get a basic directory listing with limited functionality.

2) It violates the concept that most people have of the internet. The URL doesn’t end in .com or another top level domain, and unless you link to it, n0 one will ever know it exists.

3) While the Mongoose webserver is fairly secure, this should be run on a computer isolated from your regular network.  I secured mine by limiting access only to the 192.168.* addresses provided within my test network, it’s also running on a test server outside the regular network I usually work on.

I’ll talk about more attractive methods of serving up eBooks in my next post.

 

 

Linking to Reserve Detail Screens in VTLS Virtua

A tip of the hat to my colleague in New York, she comes up with some of the best questions and gets me to explore Virtua in more detail.

Here is how to link to a reserve detail screen in VTLS Virtua.

First, I’ll show you how I figured this out.

Here is a unaltered link to an individual instructor in Virtua

http://yulib.mc.yu.edu:8000/cgi-bin/gw/yulis?sessionid=20110410003655344304&skin=portal&lng=en&inst=consortium&host=localhost%2b2222%2bDEFAULT&patronhost=localhost%202222%20DEFAULT&search=RESERVEINTR&function=RESERVE_INTR&sourcescreen=INITREQ&pos=1&rootsearch=RESERVESCAN&elementcount=1&u1=2015&t1=4143

The majority of the detail on this link can be removed. The most important parts are

http://yulib.mc.yu.edu:8000/cgi-bin/gw/yulis?search=RESERVEINTR&function=RESERVE_INTR&rootsearch=RESERVESCAN&u1=2015&t1=4143

the search=, function=, rootsearch=, u1=, and t1= portions.

This search leads to a page where you can see the individual faculty member. As we see from t1=, they are identified by an individual number, not by their name.

Here’s the fun part. Virtua seems to use the cache to link to the detail screen. It’s not clear that the Reserve Detail URL is a search of any kind at all.

An unaltered link to a detail screen looks like this.

http://yulib.mc.yu.edu:8000/cgi-bin/gw/yulis?sessionid=201104100119305353666&skin=portal&lng=en&inst=consortium&host=localhost%2b2222%2bDEFAULT&patronhost=localhost%202222%20DEFAULT&search=RESERVEINTR&searchid=2&function=RESERVEDETAIL&sourcescreen=RESERVE_INTR&pos=1&totalitems=1&itempos=1&rootsearch=RESERVESCAN

Here’s the difference: function=RESERVEDETAIL

So all we do is switch out the fuction=RESERVESCAN element on the first search with function=RESERVEDETAIL

http://yulib.mc.yu.edu:8000/cgi-bin/gw/yulis?search=RESERVEINTR&function=RESERVEDETAIL&rootsearch=RESERVESCAN&u1=2015&t1=4143

To create additional search strings for individual faculty members, you can do a search in the ‘Reserve Search’ search box for the faculty member. Copy the URL into Notepad and find the t1= number for that faculty member. Then using the base URL

http://<Your Virtua System Here>?search=RESERVEINTR&function=RESERVEDETAIL&rootsearch=RESERVESCAN&u1=2015&t1=

you can plug any faculty identifying number on the end, leading to a direct link to a list of that faculty members reserve items.

You can link to a specific course as well

http://yulib.mc.yu.edu:8000/cgi-bin/gw/yulis?search=RESERVEINTR&function=RESERVEDETAIL&rootsearch=RESERVESCAN&u1=2013&t1=chem%201377H