Evolution of a Web Site: The Virtual Gramophon Story

Search
CAML Review
Conference
Mailing List
Membership
Executive
Links
Directory

v. 32 no. 3 November / novembre 2004

Evolution of a Web Site: The Virtual Gramophon Story

By Richard Green
Library and Archives Canada

If you have attended any CAML, IASA, ARSC, or Canadian Collectors' Congress meetings in recent years, and have an interest in discography and Web sites, then you may have heard a version of this paper. It is something I regard as a work in progress and subject to continual revision and updating because, in my opinion, when it comes to the Internet, we have few definite facts nor can we accurately predict how it is going to evolve, especially with respect to libraries and archives. Institutions like Library and Archives Canada have a crucial role in determining not only what information is on the Web, but whether the Web itself will be used as a vehicle primarily for commercial purposes, as an entertainment resource, as a place to find current information, or as a means to assist scholarship and serious research. The experience of The Virtual Gramophone: Canadian Historical Sound Recordings Web site illustrates how the Web has evolved in just a few short years.

The Virtual Gramophone was one of the first fully integrated multimedia sites to get onto the Net. It was launched in the fall of 1998 at the IASA conference in Paris. CAML members may recall a preview during the summer conference in Ottawa that year. When the site was first conceived, it was envisioned as being an update of the book, Roll Back The Years. Roll Back was written in 1975 by the first head of the Library's Recorded Sound and Video Collection, Edward Moogk. In fact, it was Moogk's collection of 78-rpm discs, which he donated in 1967, that started the Sound Collection at the National Library. Roll Back the Years covered the history of recorded sound in Canada up to 1930. It was and still is the standard reference work on the beginnings of the recorded sound industry in Canada. Researchers often asked when a revised edition would appear or even a second volume to complete the history of the 78-rpm era in Canada. Though the project was endorsed by the Bibliographic Society and appeared regularly as one of the Library's goals, financial constraints made the prospect unlikely.

In the 1990s the Internet came along and, as the Library began to make information available on the Web, the Library's Technical Services staff wanted to do a project to test this new technology. The Recorded Sound and Video Collection proposed updating Roll Back the Years in a way that would take full advantage of the possibilities offered by the Web. Instead of just listing recordings, there would be images and audio files, so people could listen to the recordings as well. There was interest and, most importantly, there was money, so we were on our way. The technology people were somewhat surprised to discover what they thought of as a test would instead be on the Web and available to the public. But, after some discussion, the Virtual Gramophone was launched in November 1998.

Like Roll Back the Years, the Virtual Gramophone has a table of contents, biographies of prominent recording artists, a history of the industry, and a bibliography: in other words, the usual format of a book with the addition of computer technology. Replacing the lists of record releases and discographies of individual artists was a searchable database with twenty-seven different fields providing a wide variety of search options. The range of fields was designed to appeal to both the discographic specialist and the general public. Many entries in the database linked to a fully restored digital audio file of a recording. Rather than trying to describe the music and the styles of the time, visitors to the Virtual Gramophone could listen to the records themselves.

What wasn't anticipated, and something which many institutions including Library and Archives Canada had a hard time grasping, was the difference between a Web site and a book. The reality is that people don't come at a Web site from the front. They don't necessarily see the title information splashed across each page, nor do they look at the introduction, table of contents and index. Rather, they explore, use the hyperlinks, jump from one selection to another, have fun and let serendipity take its course. They do use the site, but not in the way that is expected in a traditional library or archive culture.

Our experience has been that people come at the Virtual Gramophone site from every direction possible through search engines like Google, through links from other pages, by entering random terms into the URL, and often just by chance. An example was the person who was searching for anything on the family name of Gerson. As a result, the searcher came to our site and found Albert Gerson, the grandfather. There was nothing in the Library's files about Albert Gerson and Roll Back only stated he had recorded two sides for His Master's Voice. Thanks to information from the family, we now know he played in theatre orchestras in Montreal and eventually moved to the United States. Many other families have also contacted the LAC. The point is that if the Gerson family had been looking in any place other than the Internet, it would have taken them years to find the one-line, un-indexed reference in the original book. It is also validates the approach taken by the Virtual Gramophone, namely to be comprehensive and include unknowns such as Albert Gerson, rather than being more selective and choosing only the best-known songs. This ongoing debate is taking place in many institutions. Is the purpose of a Web site to provide a window on the collection which encourages the visitor to enter and explore further through traditional means, or is it to give access to the complete collection?

When a book is published the author and the publisher usually have an idea of their market and the reader's level of sophistication. In the planning stages, the Virtual Gramophone was considered to be something of limited interest to people such as collectors, discographers and other serious researchers. This idea was wrong. It turns out the Virtual Gramophone is of interest to everyone. Music is a universal language and the older recordings appeal to a wide audience. Many visitors don't realize that the Virtual Gramophone is a library or even from Canada. Some believed that it was an e-store and the items were for sale!

Another difference from a book is, that when you look at a book, you are looking at a finite resource. The publishing date and the information it contains will not change until there is a new edition. A Web site is not like that nor can it be. Web sites are about change. Visitors want Web sites to evolve, they want to see new information and they want it to be current. A good illustration of that is our "Resources" page. When the site was first discussed we tried to include every possible concept. Beyond a bibliography of published sources it was felt that a page of links to sites dealing with 78-rpm and cylinder recordings would be a great idea. In this case, the old adage about being careful what you wish for definitely applies. The first version of the site provided twenty-nine quality links. In 2001, we updated the page and without much effort came up with over 350 sites, not counting the ones that were not-so-good or peripherally related, and selected 125. Each of the entries had a caption written for it that was then translated. In the years since, some of the links have changed or even disappeared. Another update is due and there will be even more sites to add. Keeping the content current is a challenge.

But content is only one aspect. A Web site is by its nature driven by rapidly changing technology. The most popular aspect of Virtual Gramophone are the audio files. These are complete, digitally restored, recordings. In 1998 RealAudio, using streaming technology, was the best available. Since then, peer-to-peer file sharing using MP3 technology has come to be preferred because it offers the possibility of downloading the file to your own system. The Virtual Gramophone added MP3 files in 2003.

A Web site like the Virtual Gramophone is a long-term commitment. It can't be put on the Web and left to gather dust like a book on the shelf. Its functionality, design and content need to be revisited on a regular basis.

Designing a site for every level of user is another challenge. The page where users accessed the online catalogue was labeled "Database." Many people didn't know what the term meant. But even if they did, they would get to the database page and then not know how to do a search page. Partly it is a question of terminology, partly it is a question of the functionality, and partly it is the user's Internet experience. Just because everyone can access the Web doesn't mean they have the same level of expertise. To correct these and other problems, a redesigned site was launched in January 2004. We changed "Database" to "Collection Search." We also changed "Audio," the page where visitors could scan the list of the audio files without searching the database, to "Listen." Some of our regular visitors asked why we had removed the audio files. The transition was there to be read on the "What's New" page, but we find that few consult the "Help" pages even if they are just a click away. Visitors would rather try on their own a few times and then give up or send cryptic e-mails to the "Comments" mailbox. Conspicuously missing from the Web is a librarian or archivist to guide the inexperienced researcher through the collection. We do it all the time for on-site visitors to our institutions. We now have to imagine how it can be done on the Web.

When a researcher visits Library and Archives Canada, a staff member can explain why some items are listed in the database and others are not. Almost every institution has its quirks in the way the collections and policies have evolved that only staff can explain. Here are some Virtual Gramophone questions we have received:

Why don't you have your whole collection available? Only so much material can be processed at a time and it is always a question of priorities and funding. Web sites like the Virtual Gramophone require a lot of money. Does anyone want to share their budget or know of a wealthy donor?
How can you call yourself a Canadian site and not have a Wilf Carter record? In short, because Wilf Carter died in 1996. He lived a rich, full, long life, which also means his songs, even the early ones that are now over seventy years old, are still under copyright and will be until January 1, 2047.
Why don't you arrange to clear copyright? We have tried. The head of a Canadian copyright collective once told me that there were only about five songs written before 1920 that have any real commercial value. But this didn't translate into a willingness to negotiate a reasonable fee when the time came to license all the other songs.

It also is a truism to say that if there is a mistake on the Web site someone will find it. Our biography of May Irwin was based on the American Grove Dictionary which, although a good source, does contain many errors, as one of her descendents pointed out.

Needless to say, some people are never happy, such as the man who complained that our cataloguing records were too detailed, that it was done by bureaucrats with too much time on their hands, and that it was a waste of taxpayers' money.

But for every negative there have been dozens of positives. In January 1999, there were about 1,200 entries in the database and 205 Real Audio files available. The first month that a statistics report was generated showed that there were 1,423 page requests and that 517 audio files were listened to. Pretty good!

Four years later, in January 2003 we had 97,368 page views which, in the Brave New World of Web stats, means we had 26,356 visits and 73, 802 audio files consulted. In 2003, there were over 600,000 visits and 1,000,000 audio files accessed. The average visitor to the site stays for twenty minutes. This was a great surprise to the people at our institution who thought the Virtual Gramophone would have some initial interest that would then die off.

From another perspective, in the 1990s on average about 1,000 78s per year were retrieved on-site from the collection for client requests or to answer reference questions. For the month of March 2003, over 100,000 songs were listened to on the Internet, which is the equivalent of 100 years' worth of on-site retrievals. If we actually had received these requests at our institution for copies, it would have taken our staff, assuming they never took a holiday, almost seventeen years. Needless to say, the Internet has had a huge impact,on improving accessibility.

The Virtual Gramophone is a rewarding site to work on. The positive response has been overwhelming from people who have found songs that they associate with relatives, family occasions or particular memories. People have enjoyed the background information, the photographs--in fact, the total package. It has been used by museums, TV and film production companies, for CD-ROMs, on radio broadcasts, by musicology and music performance students, and in schools and for educational projects. As part of the redesign, there is a new component, entitled "Educational Resources," to assist students and teachers.

It has also been interesting to watch the people hired to do most of this work evolve into passionate researchers and 78-rpm aficionados. No one has ever really done a systematic examination of Canadian recordings before. These employees are finding new information and making it available to all.

There are thousands more recordings to be added to the Virtual Gramophone. From small beginnings the database has grown to include over 12,000 titles and 3,600 audio files. In 2004 and 2005 the Virtual Gramophone will complete work on Canadian classical artists from the 78-rpm era with several new articles and additional recordings, and will begin work on a comprehensive presentation of Columbia recordings in Canada. The coming year will also see a study undertaken to move The Virtual Gramophone from a standalone database to the Library and Archives Canada's AMICUS system. This will make the cataloguing records available to all AMICUS users in the library community. You are invited to visit the Virtual Gramophone (www.collectionscanada.ca/gramophone/). Comments are welcome.


	Search CAML Review Conference Mailing List Membership Executive Links Directory		v. 32 no. 3 November / novembre 2004 Evolution of a Web Site: The Virtual Gramophon Story www.collections canada.ca/gramophone/ By Richard Green Library and Archives Canada If you have attended any CAML, IASA, ARSC, or Canadian Collectors' Congress meetings in recent years, and have an interest in discography and Web sites, then you may have heard a version of this paper. It is something I regard as a work in progress and subject to continual revision and updating because, in my opinion, when it comes to the Internet, we have few definite facts nor can we accurately predict how it is going to evolve, especially with respect to libraries and archives. Institutions like Library and Archives Canada have a crucial role in determining not only what information is on the Web, but whether the Web itself will be used as a vehicle primarily for commercial purposes, as an entertainment resource, as a place to find current information, or as a means to assist scholarship and serious research. The experience of The Virtual Gramophone: Canadian Historical Sound Recordings Web site illustrates how the Web has evolved in just a few short years. The Virtual Gramophone was one of the first fully integrated multimedia sites to get onto the Net. It was launched in the fall of 1998 at the IASA conference in Paris. CAML members may recall a preview during the summer conference in Ottawa that year. When the site was first conceived, it was envisioned as being an update of the book, Roll Back The Years. Roll Back was written in 1975 by the first head of the Library's Recorded Sound and Video Collection, Edward Moogk. In fact, it was Moogk's collection of 78-rpm discs, which he donated in 1967, that started the Sound Collection at the National Library. Roll Back the Years covered the history of recorded sound in Canada up to 1930. It was and still is the standard reference work on the beginnings of the recorded sound industry in Canada. Researchers often asked when a revised edition would appear or even a second volume to complete the history of the 78-rpm era in Canada. Though the project was endorsed by the Bibliographic Society and appeared regularly as one of the Library's goals, financial constraints made the prospect unlikely. In the 1990s the Internet came along and, as the Library began to make information available on the Web, the Library's Technical Services staff wanted to do a project to test this new technology. The Recorded Sound and Video Collection proposed updating Roll Back the Years in a way that would take full advantage of the possibilities offered by the Web. Instead of just listing recordings, there would be images and audio files, so people could listen to the recordings as well. There was interest and, most importantly, there was money, so we were on our way. The technology people were somewhat surprised to discover what they thought of as a test would instead be on the Web and available to the public. But, after some discussion, the Virtual Gramophone was launched in November 1998. Like Roll Back the Years, the Virtual Gramophone has a table of contents, biographies of prominent recording artists, a history of the industry, and a bibliography: in other words, the usual format of a book with the addition of computer technology. Replacing the lists of record releases and discographies of individual artists was a searchable database with twenty-seven different fields providing a wide variety of search options. The range of fields was designed to appeal to both the discographic specialist and the general public. Many entries in the database linked to a fully restored digital audio file of a recording. Rather than trying to describe the music and the styles of the time, visitors to the Virtual Gramophone could listen to the records themselves. What wasn't anticipated, and something which many institutions including Library and Archives Canada had a hard time grasping, was the difference between a Web site and a book. The reality is that people don't come at a Web site from the front. They don't necessarily see the title information splashed across each page, nor do they look at the introduction, table of contents and index. Rather, they explore, use the hyperlinks, jump from one selection to another, have fun and let serendipity take its course. They do use the site, but not in the way that is expected in a traditional library or archive culture. Our experience has been that people come at the Virtual Gramophone site from every direction possible through search engines like Google, through links from other pages, by entering random terms into the URL, and often just by chance. An example was the person who was searching for anything on the family name of Gerson. As a result, the searcher came to our site and found Albert Gerson, the grandfather. There was nothing in the Library's files about Albert Gerson and Roll Back only stated he had recorded two sides for His Master's Voice. Thanks to information from the family, we now know he played in theatre orchestras in Montreal and eventually moved to the United States. Many other families have also contacted the LAC. The point is that if the Gerson family had been looking in any place other than the Internet, it would have taken them years to find the one-line, un-indexed reference in the original book. It is also validates the approach taken by the Virtual Gramophone, namely to be comprehensive and include unknowns such as Albert Gerson, rather than being more selective and choosing only the best-known songs. This ongoing debate is taking place in many institutions. Is the purpose of a Web site to provide a window on the collection which encourages the visitor to enter and explore further through traditional means, or is it to give access to the complete collection? When a book is published the author and the publisher usually have an idea of their market and the reader's level of sophistication. In the planning stages, the Virtual Gramophone was considered to be something of limited interest to people such as collectors, discographers and other serious researchers. This idea was wrong. It turns out the Virtual Gramophone is of interest to everyone. Music is a universal language and the older recordings appeal to a wide audience. Many visitors don't realize that the Virtual Gramophone is a library or even from Canada. Some believed that it was an e-store and the items were for sale! Another difference from a book is, that when you look at a book, you are looking at a finite resource. The publishing date and the information it contains will not change until there is a new edition. A Web site is not like that nor can it be. Web sites are about change. Visitors want Web sites to evolve, they want to see new information and they want it to be current. A good illustration of that is our "Resources" page. When the site was first discussed we tried to include every possible concept. Beyond a bibliography of published sources it was felt that a page of links to sites dealing with 78-rpm and cylinder recordings would be a great idea. In this case, the old adage about being careful what you wish for definitely applies. The first version of the site provided twenty-nine quality links. In 2001, we updated the page and without much effort came up with over 350 sites, not counting the ones that were not-so-good or peripherally related, and selected 125. Each of the entries had a caption written for it that was then translated. In the years since, some of the links have changed or even disappeared. Another update is due and there will be even more sites to add. Keeping the content current is a challenge. But content is only one aspect. A Web site is by its nature driven by rapidly changing technology. The most popular aspect of Virtual Gramophone are the audio files. These are complete, digitally restored, recordings. In 1998 RealAudio, using streaming technology, was the best available. Since then, peer-to-peer file sharing using MP3 technology has come to be preferred because it offers the possibility of downloading the file to your own system. The Virtual Gramophone added MP3 files in 2003. A Web site like the Virtual Gramophone is a long-term commitment. It can't be put on the Web and left to gather dust like a book on the shelf. Its functionality, design and content need to be revisited on a regular basis. Designing a site for every level of user is another challenge. The page where users accessed the online catalogue was labeled "Database." Many people didn't know what the term meant. But even if they did, they would get to the database page and then not know how to do a search page. Partly it is a question of terminology, partly it is a question of the functionality, and partly it is the user's Internet experience. Just because everyone can access the Web doesn't mean they have the same level of expertise. To correct these and other problems, a redesigned site was launched in January 2004. We changed "Database" to "Collection Search." We also changed "Audio," the page where visitors could scan the list of the audio files without searching the database, to "Listen." Some of our regular visitors asked why we had removed the audio files. The transition was there to be read on the "What's New" page, but we find that few consult the "Help" pages even if they are just a click away. Visitors would rather try on their own a few times and then give up or send cryptic e-mails to the "Comments" mailbox. Conspicuously missing from the Web is a librarian or archivist to guide the inexperienced researcher through the collection. We do it all the time for on-site visitors to our institutions. We now have to imagine how it can be done on the Web. When a researcher visits Library and Archives Canada, a staff member can explain why some items are listed in the database and others are not. Almost every institution has its quirks in the way the collections and policies have evolved that only staff can explain. Here are some Virtual Gramophone questions we have received: Why don't you have your whole collection available? Only so much material can be processed at a time and it is always a question of priorities and funding. Web sites like the Virtual Gramophone require a lot of money. Does anyone want to share their budget or know of a wealthy donor? How can you call yourself a Canadian site and not have a Wilf Carter record? In short, because Wilf Carter died in 1996. He lived a rich, full, long life, which also means his songs, even the early ones that are now over seventy years old, are still under copyright and will be until January 1, 2047. Why don't you arrange to clear copyright? We have tried. The head of a Canadian copyright collective once told me that there were only about five songs written before 1920 that have any real commercial value. But this didn't translate into a willingness to negotiate a reasonable fee when the time came to license all the other songs. It also is a truism to say that if there is a mistake on the Web site someone will find it. Our biography of May Irwin was based on the American Grove Dictionary which, although a good source, does contain many errors, as one of her descendents pointed out. Needless to say, some people are never happy, such as the man who complained that our cataloguing records were too detailed, that it was done by bureaucrats with too much time on their hands, and that it was a waste of taxpayers' money. But for every negative there have been dozens of positives. In January 1999, there were about 1,200 entries in the database and 205 Real Audio files available. The first month that a statistics report was generated showed that there were 1,423 page requests and that 517 audio files were listened to. Pretty good! Four years later, in January 2003 we had 97,368 page views which, in the Brave New World of Web stats, means we had 26,356 visits and 73, 802 audio files consulted. In 2003, there were over 600,000 visits and 1,000,000 audio files accessed. The average visitor to the site stays for twenty minutes. This was a great surprise to the people at our institution who thought the Virtual Gramophone would have some initial interest that would then die off. From another perspective, in the 1990s on average about 1,000 78s per year were retrieved on-site from the collection for client requests or to answer reference questions. For the month of March 2003, over 100,000 songs were listened to on the Internet, which is the equivalent of 100 years' worth of on-site retrievals. If we actually had received these requests at our institution for copies, it would have taken our staff, assuming they never took a holiday, almost seventeen years. Needless to say, the Internet has had a huge impact,on improving accessibility. The Virtual Gramophone is a rewarding site to work on. The positive response has been overwhelming from people who have found songs that they associate with relatives, family occasions or particular memories. People have enjoyed the background information, the photographs--in fact, the total package. It has been used by museums, TV and film production companies, for CD-ROMs, on radio broadcasts, by musicology and music performance students, and in schools and for educational projects. As part of the redesign, there is a new component, entitled "Educational Resources," to assist students and teachers. It has also been interesting to watch the people hired to do most of this work evolve into passionate researchers and 78-rpm aficionados. No one has ever really done a systematic examination of Canadian recordings before. These employees are finding new information and making it available to all. There are thousands more recordings to be added to the Virtual Gramophone. From small beginnings the database has grown to include over 12,000 titles and 3,600 audio files. In 2004 and 2005 the Virtual Gramophone will complete work on Canadian classical artists from the 78-rpm era with several new articles and additional recordings, and will begin work on a comprehensive presentation of Columbia recordings in Canada. The coming year will also see a study undertaken to move The Virtual Gramophone from a standalone database to the Library and Archives Canada's AMICUS system. This will make the cataloguing records available to all AMICUS users in the library community. You are invited to visit the Virtual Gramophone (www.collectionscanada.ca/gramophone/). Comments are welcome.