Peter MacDonald | 1 Jun 2004 20:12

[Fedora-users] A Fedora ingest procedure

If you need some ideas on how to go about ingesting records into Fedora, 
the following shows you how I did it. It's not pretty, but it works.

The techniques, files, and scripts mentioned below were designed to be used 
in processing Hamilton College's collection of Henry Welch letters 
(American Civil War era), so for it to work for you you will need to 
customize it significantly.

Here goes:

According to the Fedora instructions you can use the supplied 
mets-template.xml file (you need to customize it though) as a pattern for 
ingesting mets-fedora objects, but the mets-template and the mets-fedora 
objects being ingested by them have to match in certain ways or the ingest 
will fail. My problem was that our objects have many different numbers 
of  datastreams -- most are between 1 and 10 images, but one consists of 71 
images, thus requiring 142 datastreams if you make JPEGs and a thumbnail. 
This situation required a different mets-template for each batch of objects 
that would need the same number of datastreams.

- I created a series of mets-templates (based on examples supplied with 
Fedora): one to be used for objects consisting of a single datastream (say, 
a pdf file), and a template for objects with 2 datastreams (say, a JPEG and 
a text file), and so on (e.g., I created a template that could accommodated 
an object consisting of 1 JPEG image and 1 GIF thumbnail, and one that 
could accommodate 2 JPEGs and 2 GIF thumbnail images, one that could 
accommodate 3 JPEGs, 3 GIF thumbnails, and 1 TEI/text file, and so on until 
I had about 15 different templates, named like: mets-template_2images.xml, 
mets-template_3images.xml, and so. For an example of one of these templates 
look at:
(Continue reading)

Thornton Staples | 11 Jun 2004 19:45
Favicon

[Fedora-users] Fedora funding approved

  I am very please to announce that our proposal to the Mellon Foundation
for another three years of funding for the Fedora project was approved 
by the board this morning. This grant, for $1.4 millon will allow us to 
add another programmer to the team and greatly extend the functionality 
of the software. We will post the text of the proposal on our website 
next week.

Thorny

Woodall, Nancy | 15 Jun 2004 17:13

[Fedora-users] Help with server specs

The University of Richmond plans to use FEDORA as a repository for
digitized Civil War newspapers.  This is part of an IMLS grant that we
share with Tufts.  The grant specifies 1.4 T for images. At this point,
we plan to use Oracle 9i as our RDBMS. Has anyone tried a more recent
Oracle release?  We will use Red Hat Linux for OS. 

We would really appreciate any advice or comments on our current specs
(see below).

We plan to add 2 72 or 146 GB SCSI drives. We intend to store the images
on a SAN with a Fiber Channel connection to the FEDORA server. Anyone
know if that could cause problems?

ProLiant DL580 G2 Intel(r) Xeon(tm) Processor MP at 2.70GHz/2MB (2P
Model)
Two Intel(r) Xeon(tm) Processors MP 2.70GHz/2MB 
Intel(r) Xeon(tm) Processor MP 2.70GHz/2MB - Option Kit
4GB Base Memory 4x512, 4x512
Standard One Ultra3 SCSI Drive Cage (2x2 Duplex std or 4x1 Simplex)
Integrated Smart Array 5i Plus Controller (Dual Channel, Ultra3)
Standard Battery Backed Write Cache Enabler (up to 64MB Write Cache)
RAID 0 setting (Requires minimum of 2 matching drives)
36.4GB Pluggable Ultra320 SCSI 10,000 rpm Universal Hard Drive (1")
36.4GB Pluggable Ultra320 SCSI 10,000 rpm Universal Hard Drive (1")
1.44MB Floppy Disk Drive 
24X IDE Slim-line CD-ROM
HP NC7170 Dual Port PCI-X 1000T Gigabit Server Adapter
SmartStart & Insight Manager
Two (2) 800W Hot Plug Redundant Power Supplies
Redundant Hot Plug Fans 
(Continue reading)

Jason Simms | 23 Jun 2004 20:22
Picon
Favicon

[Fedora-users] Fedora Frontend

Hello,

I am wondering whether much progress has been made towards a viable 
frontend to Fedora.  It seems that several institutions are using 
Fedora, and other than the supplied administrator's interface, I am 
curious what tools people are using to access, manage, and browse 
collections.  Are people using "home brew" tools?  Browser-based solutions?

I have installed Fedora, but have done nothing further than that.  I 
have quite a bit of experience with DSpace, and of course that system 
provides a very nice built-in browser interface, but the limitations 
with metadata, etc. are somewhat frustrating.  I would like to move 
towards a Fedora-based solution, but I am disappointed in the access 
tools available.

Any suggestions would be much appreciated.  Thanks!

Jason Simms

LeVan,Ralph | 23 Jun 2004 20:53
Picon
Favicon

RE: [Fedora-users] Fedora Frontend

Why not DSpace on top of Fedora?

Ralph

> -----Original Message-----
> From: Jason Simms [mailto:jsimms@...]
> Sent: Wednesday, June 23, 2004 2:23 PM
> To: fedora-users@...
> Subject: [Fedora-users] Fedora Frontend
> 
> Hello,
> 
> I am wondering whether much progress has been made towards a viable
> frontend to Fedora.  It seems that several institutions are using
> Fedora, and other than the supplied administrator's interface, I am
> curious what tools people are using to access, manage, and browse
> collections.  Are people using "home brew" tools?  Browser-based
> solutions?
> 
> I have installed Fedora, but have done nothing further than that.  I
> have quite a bit of experience with DSpace, and of course that system
> provides a very nice built-in browser interface, but the limitations
> with metadata, etc. are somewhat frustrating.  I would like to move
> towards a Fedora-based solution, but I am disappointed in the access
> tools available.
> 
> Any suggestions would be much appreciated.  Thanks!
> 
> Jason Simms
> 
(Continue reading)

Jason Simms | 23 Jun 2004 20:59
Picon
Favicon

Re: [Fedora-users] Fedora Frontend

That is an interesting idea, but I don't think it can be done (easily). 
  Much of the Java code for DSpace is laced with information specific to 
DSpace; that is, they do not do a good job of separating form from 
function.  For instance, they do not abstract the database layer (or 
authentication, etc.), so code specific to the PostgreSQL database, data 
model, etc. is right alongside the business logic.  Very frustrating.

I would need to rewrite the vast majority of the code, I would think, 
unless you know of a way that you or another user has successfully 
integrated Fedora with a DSpace frontend.

Jason Simms

LeVan,Ralph wrote:
> Why not DSpace on top of Fedora?
> 
> Ralph
> 
> 
>>-----Original Message-----
>>From: Jason Simms [mailto:jsimms@...]
>>Sent: Wednesday, June 23, 2004 2:23 PM
>>To: fedora-users@...
>>Subject: [Fedora-users] Fedora Frontend
>>
>>Hello,
>>
>>I am wondering whether much progress has been made towards a viable
>>frontend to Fedora.  It seems that several institutions are using
>>Fedora, and other than the supplied administrator's interface, I am
(Continue reading)

Sandy Payette | 23 Jun 2004 21:24
Picon
Favicon

RE: [Fedora-users] Fedora Frontend

To date, institutions have built their own front-ends to Fedora, using the
Fedora web service APIs to make appropriate calls to the Fedora repository.
Good examples of this practice include the University of Virginia,
Northwestern, Tufts, the NSDL, plus others.  Interestingly, many institutions
that have adopted Fedora consider it a benefit that Fedora has not fused the
end-user experience into the repository service, and are happy for the open
APIs which enable custom workflow and user interfaces to be built on top of
Fedora.

That said, the Fedora project will provide more in the way of
tools/utilities/UIs for object creation and workflow.  We will also provide a
configurable web interface for simple content submission.  This will occur in
Phase 2 which has recently been funded.  The Fedora project will receive $1.4
million from the Mellon foundation over the next three years.  One major
focus points of the Phase 2 project will be object creation tools and
workflow.  As with Phase 1, we do not want to assume one particular workflow,
so we will be working with several of our collaborators to create an
adaptable and configurable workflow solution, based on current workflow
standards.  The text of the Phase 2 proposal will be available on the Fedora
web site soon (www.fedora.info).

Phase 2 will begin this October.

- Sandy

> -----Original Message-----
> From: fedora-users-bounces@... 
> [mailto:fedora-users-bounces@...] On Behalf Of Jason Simms
> Sent: Wednesday, June 23, 2004 3:00 PM
> To: LeVan,Ralph
(Continue reading)

John Fereira | 23 Jun 2004 21:29
Picon
Favicon

Re: [Fedora-users] Fedora Frontend

At 02:59 PM 6/23/2004 -0400, Jason Simms wrote:
>That is an interesting idea, but I don't think it can be done 
>(easily).  Much of the Java code for DSpace is laced with information 
>specific to DSpace; that is, they do not do a good job of separating form 
>from function.  For instance, they do not abstract the database layer (or 
>authentication, etc.), so code specific to the PostgreSQL database, data 
>model, etc. is right alongside the business logic.  Very frustrating.

That was the major problem that I had with DSpace when I first started 
looking at it.  It's been awhile since I looked at it but I was under the 
impression that DSpace was doing some work to integrate Fedora into it.

I think it's important to keep in mind that unlike DSpace, Fedora doesn't 
try to be a complete library in a box.  I think of Fedora as the "database" 
layer, or to be more accurate, the persistence layer.

I am currently refactoring a system which has been around in our library 
for over five years.  It currently uses a database backend and perl cgis 
for access to the database.  It's essentially a document repository.

For the redesign, I am using the Spring Framework (www.springframework.org) 
using the Spring controllers and jsp pages (with JSTL) for the presentation 
layer.  Rather than try to do complex searching/browsing of the collection 
using the Fedora repository, I'm continuing to use a mysql backend database 
for document/search info.  Using the Spring Framework, all of the business 
objects are implemented using DAOs (Data Access Objects) with iBatis 
sqlMaps for access to the persistence layer.  What I am hoping to do is 
access the actual documents (which are available in multiple formats) with 
a DAO implementation as a web service client accessing a Fedora repository.

(Continue reading)

Jason Simms | 23 Jun 2004 21:36
Picon
Favicon

Re: [Fedora-users] Fedora Frontend

Sandy,

Thank you for your prompt and thorough reply.  I agree with the decision 
to separate the repository from the interface.  However, I also 
understand that any institution wishing to use Fedora will probably want 
*some* type of interface.

Building an interface does obviously necessitate some agreement on a 
base set of features that many institutions can agree are necessary. 
Doing so also means that any interface agreed upon will not necessarily 
meet exactly the needs of every institution, thus requiring further 
customization.

I believe, however, that a happy medium probably exists.  I agree that a 
rigid interface need not be constructed, but I also think that a 
boilerplate interface would be welcomed.  To this end, I am pleased to 
hear about Phase 2 of Fedora, and I am glad for the continued funding. 
I am just trying to avoid reinventing the wheel, so to speak.  If ten 
institutions have already built custom interfaces, perhaps some of them 
would be willing to provide their code as a base from which to work.

If the sense is that this is impossible, then we at the University of 
Tennessee will work towards a custom-built solution of our own, but we 
also want to pursue collaborative efforts whenever possible to avoid 
wasted time.  Thanks!

Jason Simms

Sandy Payette wrote:

(Continue reading)

Jason Simms | 23 Jun 2004 21:46
Picon
Favicon

Re: [Fedora-users] Fedora Frontend

I like this type of idea very much.  I believe in separating the 
components as much as possible so that the best tool can be integrated 
for each service.  That is why I am anxious to implement Fedora here 
before we move too far down the road with DSpace (although I imagine it 
wold be possible to export from DSpace and import into Fedora with a 
little elbow grease!).

Is your work extensible at all, or will be linked at the hip with your 
specific setup?  The reason I ask is that I envisioned something like a 
Perl/PHP/JSP/etc. frontend interfacing with Fedora, and your design 
seems elegant.  In short - if your solution could be adapted to our 
needs, we would be happy to collaborate, test, document, etc. as needed. 
  Thanks!

Jason Simms

John Fereira wrote:

> At 02:59 PM 6/23/2004 -0400, Jason Simms wrote:
> 
>> That is an interesting idea, but I don't think it can be done 
>> (easily).  Much of the Java code for DSpace is laced with information 
>> specific to DSpace; that is, they do not do a good job of separating 
>> form from function.  For instance, they do not abstract the database 
>> layer (or authentication, etc.), so code specific to the PostgreSQL 
>> database, data model, etc. is right alongside the business logic.  
>> Very frustrating.
> 
> 
> That was the major problem that I had with DSpace when I first started 
(Continue reading)


Gmane