Incredible - Duplicate is altering/overwriting its own source!

Klaus Wendel

unread,

Jul 16, 2010, 6:04:24 AM7/16/10

to ICA-AtoM Users

Hallo,

If you are duplicating an archival description, the SOURCE-object is
loosing any information of the date-field! That's really malicious,
since normally you won't have a look at an already finished set of
data, if you are just beginning a new one.

By now, several times I bite back my comments, because as an open
source user, I HAVE to be very modest and polite on critics and as an
(open source) developer as well, I HAVE to be sensible for bugs: They
simply happen!
But this time, I cannot retain myself, since this bug must result from
the very strange data model, which made me a lot of headache nine
months ago, as I was forced to remove accidentally applied access
points manually with phpmyadmin. No wonder, that you are preferring
the Apache Lucene search engine, building index files on disk, instead
of the very much more powerful built-in query language of SQL.

The conception of ICA-Atom with repositories and physical objects is
really great! I have never seen this in other even commercial
products. And that's the reason for me to favour ICA-Atom. However,
will there be a real chance in future, that your good ideas will
become workable for a large amount of data? (We do not even have a
really large amount! We have serious problems with only approx. 900
archival descriptions and approx. 2600 access points.)

Now, let's get back to almost any archival description, we have
created and duplicated since yesterday and recover the date
information.

Regards,

Klaus

David Juhasz

unread,

Jul 16, 2010, 2:42:55 PM7/16/10

to ica-ato...@googlegroups.com

Hi Klaus,

You are right that this is a very serious bug, and it's a shame that we missed it in our testing for Release 1.0.9. We had actually caught and fixed the issue once before the Release process, but it was reverted in a later commit. Jesús has created a patch, which you can download from http://www.sevein.com/tmp/issue1330-icaatom109.patch

I can understand that it's frustrating and alarming to lose data when using the application, but we did not introduce this bug on purpose. We do our best to test the application thoroughly before each release, but even huge companies like Microsoft still regularly release software with critical bugs that slip through the quality assurance process. This bug has nothing to do with the underlying data model, it is simply a bug in the code.

Also, this project is not only open source, but in it's beta testing phase - we do not, nor have we ever made any guarantees that the software would be bug free, nor can we provide any guarantees against data loss; we recommend backing up your data regularly. I'd like to re-iterate that we do our best to try and provide bug free software, but we are also a small software firm providing a free (gratis) and open source product with limited resources.

If you do not like the Apache Lucene search engine you are welcome to change the application to use SQL for full-text searching. Alternatively if you are willing to fund such a modification I'm sure you could find a developer willing to make the changes. We appreciate your feedback and testing of ICA-AtoM, and we are happy to listen to your opinions about the direction the project should take, but we simply do not have resources to make the software provide everything to everyone, and we need to prioritize features that we can get funded or that we think will provide the most benefit for the time investment.

As for the scalability of the application, we feel that we have made a lot of progress in this area in the last two releases, and we plan on improving further in the future. We appreciate any detailed reports regarding the pain points in doing data entry or retrieval for large datasets and have responded to such reports from users (including some of your reports) in this release and previously.

Regards,

David Juhasz,

Software Engineer, Artefactual Systems Inc.

http://www.artefactual.com | P: 604.527.2056 | F: 604.521.2059

Klaus Wendel

unread,

Jul 21, 2010, 7:07:47 AM7/21/10

to ICA-AtoM Users

Hi David,

many thanks for the fast built patch!

Well, you are right, if you say: Don't talk, wise guy, contribute and
make it better. I know the situation very much on both sides. I have
my own project(s), where customers are getting a tool for free and are
chasing me to improve and to keep myself active. I really know your
situation and it is definitely not my intension, to make your life
even harder!
Indeed, some years ago, I was forced to build some SQL-requests for a
-- in my very own opinion -- very encapsulated set of tables, which
has been exported in parts from another SQL-database. It was a kind of
Augias' stable (but NOT from the AUGIAS Database, to prevent
misunderstandings!) and I had the choice to create a tool that
rebuilds the database routinely in a running process or to apply my
SQL-search strings to the current dataset. I decided to hold the
current structure and created a number of simplifying VIEWS. A "view"
is looking like a table but it is collecting the results of any even
very complex SQL-search clearly. It is very suitable to SIMULATE the
behavior of another search interface without the need of rewriting
code in a big way. Perhaps the approach to create several views could
be applicable here, too?

Is there actually a list and description of all available patches?

Regards, Klaus

On 16 Jul., 20:42, David Juhasz <da...@artefactual.com> wrote:
> Hi Klaus,
>
> You are right that this is a very serious bug, and it's a shame that
> we missed it in our testing for Release 1.0.9. We had actually
> caught and fixed the issue once before the Release process, but it was
> reverted in a later commit. Jesús has created a patch, which you can

> download fromhttp://www.sevein.com/tmp/issue1330-icaatom109.patch

> Software Engineer, Artefactual Systems Inc.http://www.artefactual.com| P: 604.527.2056 | F: 604.521.2059