For now, if you fill the form and then upload your file, the form is reset after the file upload, erasing all you have written.

Scénario “Vérification de la qualité des notices” :
Alain veut s’assurer de la qualité des notices en vérifiant plusieurs points: les doublons possibles, les problèmes de saisies (souvent liés aux noms d’auteurs) et le référencement (est-ce que l’article est répertorié ou non); pour ce faire il clique sur le bouton “vérifier la qualité” en bas à gauche de la plateforme. Suite à ça il obtient un filtrage des articles qui ont une qualité “suspecte”.

Tiré à part should be an OAI-PMH repository so that it could be harvested by institutional harvesters (CNRS HAL for example).

References

• http://www.oaforum.org/tutorial/
• http://www.openarchives.org/OAI/openarchivesprotocol.html
• http://www.openarchives.org/OAI/2.0/guidelines-repository.htm

.... especially for bookmarks and search engines.

Our goal is to develop a web page on which we can do all our functional tests.
This page composites a list of tests, a block of test information and a "Selenium test runner". What's more, there are already the code on javascript for integrate the "Selenium test runner". url: http://gmaps-api-issues.googlecode.com/svn/trunk/selenium/core/scripts/selenium-testrunner.js

Reproduction scenario :

• open a record
• save it
• upload an attachment

... get truncated in the form.

For example:
http://tire-a-part.test.hypertopic.org/f5d111ceb74ce0b7ece2dddcf90025d5

Some participants use external tools (such as Endnote) to create or fill in information about a publication.
When they want to switch from one tool to another, they have to type all the information again, creating un-synchronized duplicates on different applications.

We need to provide a way to synchronize data transmitted via the application or via an external tool such as Endnote.

I (successfully) deposed a paper in Tire-a-part (see : http://publications.tech-cico.fr/936342e018f7d9a708dc086707742966 ) with new authors (such as Pascal Salembier..) previously unknown in the base. But after that, the Surname-Name of these authors do not appear as usual in the search menu.

The research director need the qualification of all articles in order to evaluate these articles and make decisions for the teams.
The qualification possesses:

• the reference of article
• the platform used
• year of publication
• the axis of team thematic (axes thématiques d'équipe)
• the axis of crossing thematic (axes thématiques transversales)
• if exist, the co-publisher of the article and the container

For now, saving metadata removes the attachments from the object.

So it can be neither modified nor deleted.

The title should be mandatory.

We can create an empty record and without filling the different inputs, like article 1 and 2.

We need to find a way to check the classification (if it's classified highly) selected by the user while adding an article.
The checkout / notification mustn't block the user in the process.

Example: Lepetit, Alexandre is parsed as Alex Lepetit, re.

In our website, there is a fonction which allows users to search the articles by the author names. But here is the problem: an author sometimes changes his/her author names for certain reasons. For exemple, Mary Ford is a writer, but when she is married to Mr Burn, her name is changed to Mary Burdon then. Actuelly Mary Ford and Mary Burn are the same person, so when we search either name, all the articles of Mary should be listed. This problem can be settled when we add a personal space and set an author name.
To make things easier, we make two hypothsises:

1. There aren't two authors who share the same full name.
2. This is a internal community, which means all the authors have a account.

The project "tire-a-part" has a lot of unpublished articles stored in a directory. The problem is that nobody can't use them. The aim of this task is to make them workable.

To permit the team manager to have statistical information about articles enabling him to synthesize the work done. To assist him to analyze the evolution, understand current situation and estimate future results.

The AERES categories (or types) can change over the years, therefore we need to make a feature to manage them.

A user experienced troubles while copying and pasting a BibTeX reference from JabRef.

The converted JSON text could not be parsed. It seems that it was due to hidden characters on multi-line attributes (probably \r and maybe \t).

Workarounds:

• Either copy and paste the reference in a text editor. Then, copy and paste the "sanitized" text in the import dialog,
• Or edit the reference in the import dialog so that each attribute is on a single line.

Reproduction scenario:

• open the browser (with no active authentication),
• get a record,
• select and upload an attachment,

On Safari, the upload never ends.
On Firefox and Chrome, a connection reset exception is shown.

• To-Write : https://github.com/benel/Tire-a-part/wiki/D%C3%A9clarer-un-article-comme-publi%C3%A9 / Renseigner les informations / Indexé par
• To-Write : https://github.com/benel/Tire-a-part/wiki/Consulter-les-notices-des-articles-publi%C3%A9s / Partager le rapport d'activité (URL de filtrage)
• To-Write : check the TODOList at the end of https://github.com/benel/Tire-a-part/wiki/Int%C3%A9gration-avec-la-gestion-des-identifiants-personnels-%28ldap%29

• for an unpublished article,
• for a published article.

In BibTeX import, fields should be case-insensitive.

Example from BibDesk:

@article{Lort12a,
     Author = {A. Lorton and M. Fouladirad and A. Grall},
     Journal = {European Journal of Operational Research},
     Number = {3},
     Pages = {443--454},
     Title = {A methodology for probabilistic model-based prognosis},
     Volume = {225},
     Year = {2013}}

In the form page, saving has different effects depending on the browser:
in Safari (5.1) and Chrome (14.0) : the PUT request is "canceled" for no apparent reason.

On the contrary it works fine in Firefox (6.0.2).

... on one of the combo boxes.

We can click the button "deposer" directly without choosing a file. And it answers OK .If we do so in an existing article, we will see a file of 0Mo.

The aim is to run selenium tests on the server side.

We will be using Travis as the integration server.
XVFB will enable to run an instance of firefox with Selenium driver.

... concerning references without PDF nor abstract.

There are missing scenario and mockups here : https://github.com/benel/Tire-a-part/wiki/D%C3%A9clarer-un-article-comme-publi%C3%A9 ; check the "Renseigner les informations" and "Partager le rapport d'activité" sub-sections.

Team manager should be able to follow article lifecycle and view changes made on them in a simple way.
A tool could be implemented to help this kind of person getting a quick overview of all changes sorted by date.

We need to add a feature which allow users to add a new category or a type of article.

The year should be mandatory.

i.e. Dublin Core attributes in HTML "meta".

This inability seems to be related to the way HTML form uploads are handled on Windows rather than on a bug specific to this application.

Workaround

Remove accents from the file name... or use another operating system ;)

For now, a JSON message is shown instead of the updated form.

Workaround: Get the URL again.

... should be a comma and a space (instead of a comma).
It must be fixed both on display and save.

when you save a record, you don't get any message confirming that your record has been saved and there is no redirection to the main page