Xper2 - Use FAQ



General presentation of Xper2

What does the Xper2 interface look like?

Xper2 au démarrage

Xper2 comes up with three modules, each one with its own interface. To navigate from one module to another use the menu « Edit »:
- The sub menu “Edit by descriptors” opens the descriptive system edition module.
- The sub menu “Edit by item” opens the module of knowledge edition.
- The sub menu “Identification” opens the CAI module.

What do the Xper2 menus look like?

Four principal choices are available in Xper2  toolbar

« File »
« Open a base »: Open an existing KB
« New base »: Create a new KB
« Import -> XPER Original Format »:import a KB in XPER format
« Export -> XPER Original Format »: export a KB in XPER format
« Save »:  Save a KB
« Save as... »: Save a KB with another name
« Close the base »: Close the current KB
« Quit »: Close the current KB and quit Xper2
« Edit »
« Edit by descriptors » : Display the edition module of the descriptive system
« Edit by item » : Display the edition module of the KB
« Identification » : Display the CAI module
« Tools »
« Control the base -> Control everything » : check the integrity of the data
« Properties » : Display the properties of the KB
« Help »
« About... » : Display the About... window with the version of the program and the address of the LIS

N.B., At start, only the menus « File » and « Help » are accessible. The « Edit » menu is available in the complete version of Xper2.

How to create a new Knowledge Base (KB)?

In order to create a new KB, click on  « New base » in the menu « File ». A window will invite you to enter the name of the KB. Once entered, the Edition by descriptors module will automatically be opened. (cf. Edition of descriptors)

N.B. You can only create new KBs on the complete version of  Xper2!

How to open an existing KB?

Click on « Open an existing base » in the menu menu « File ». You will be invited to browse your hard disk in order to select the KB in a «xpd» format.

Is it possible to open several KB at the same time?

No, it is not. Xper2 can only work if a single instance of a KB. This means that you cannot open or create a KB if tjere is one already open.

How to save a KB?

If it is a new one, « Save » and « Save as » does the same. You will be invited to browse your hard disk and choose a file name. The extension (xpd) is automatically created. If you base exists already, « Save » will save your base on the same file. « Save as » will allow you to save your base under another file name.

Note, If you close your base or the software by mistake, a textbox will invite you to save your base.

Which files are created when saving a KB?

When a KB is saved in Xper2 format, 4 files are created:

  • « xpd » : the descriptive system + the list of individuals (format xml)
  • « xpg » : the groups of variables (text format) - deprecated
  • « xpr » : relationships among variables (text format)
  • « xpw » : weight (texte format) - not used for the moment
  • « xpx » : item x descriptors matrix (text format)

The folder « images » is created if it does not already exist. It is used to stock the images illustrating your KB.

N.B. when you create a KB, save it before adding imagesin order to have the « images » folder created. Otherwise, you may find some problems concerning image management.

What files are created when exporting to XPER format?

Exporting to XPER format will create 8 files:

  • « xpi » : the list of the names of individuals (text format)
  • « xpg » : the groups of variables (text format)
  • « xpm » : the list of states (text format)
  • « xpp » : the KB parameters (text format)
  • « xpr » : the relationships among variables (text format)
  • « xpv » : the list of variables (text format)
  • « xpw » : weighting scheme (text format)
  • « xpx » : the individuals x variables matrix (text format)

Is it possible to import a base written in a different format?

XPER² can import KBs written in the old XPER format, in CSV (example here) and in SDD.

Is it possible to export my knowledge base in different format?

XPER² can export KBs written in the XPER format,in CSV, SDD, HTML and NEXUS format.

How to share KBs?

Given the fact that a KB is made up of several files (cf. Which files are created when saving a KB?) and a set of images, it is advised to put all the files together into a zip file in order to abtain a single file. This solution allows you to easily email your KB while keeping a total file compatibility.

Which are the properties of a KB?

in order to access the informations of a KB, click on « Properties » of the menu « Tools ».
Proprietés
A KB has a name, a date of first edition (date of creation) a date for the last edition, a description or comment and the list of its authors. The dates are given as a mention and cannot be edited. However, you can edit the name, the description and the llist of contributors of a KB.

N.B. Dates are written after the first saving of the KB.

How to rename a KB?

Open « Properties » window (cf. Which are the properties of a KB?), and edit the field « Name ». Save your changes clicking « Submit ».

N.B. the properties file is updated when the  KB is saved frome the menu « File ».

How to edit the description of a KB?

Open the « Properties » window (cf. Which are the properties of a KB?), and edit the « Knowledge base description » field. Save your changes clicking « Submit ».

N.B. the properties file is updated when the  KB is saved frome the menu « File ».

How to edit the names of the author of a KB?

Open the « Properties » window (cf. Which are the properties of a KB?), and edit the « Knowledge base description » field.
In order to add an author, click on  Ajouter and enter his (her) name. The author's name will then appear in the list on the right side.
To remove an author, select him (her) and click on  Supprimer.
To modify an author's name, select it from the list and click on  Editer.
Modify its value and save.

N.B. the properties file is updated when the  KB is saéved frome the menu « File ».

What is the terminology used in Xper2?

Xper2 uses the usual terms applied in life sciences. Obviously, this jargon is not suited for all the possibles applications of Xper2. Note that the terminology used by XPER has been updated in Xper2. Please refer to the following correpondence table:


  XPER Xper2
Concept Individual item
Descripteur Variable descriptor
Value State descriptor-state
Descriptive system Variables + states descriptors + descriptor-states
Entity Group Group

How to know the version of Xper2?

OPen the « About... » window from the « Help » menu. On the title bar of the window you can see the name of the module (Xper2, Xper2-itemEdition, Xper2-Identification) and its version (e.g. 1.3)

How to check the integrity of the data of a KB?

On the « Tools » menu, select « Check base ».

How are checked the data?

There are three levels of control that allow you to know if your descriptive system is correctly defined (no unused descriptors, always used,...) and if your data is correctly entered (no doubles,...):

1-  descriptor control: it checks if descriptors are correctly described, i.e. there is not a misuse of child descriptor
Check of the descriptors

2- State control: checks if any state is never or always used
Check of the states

3- item control: checks if item are discriminated and if there are item with identical descriptions
Check of the items

How does the image manager work?

Image manager
The image manager allows the edition of illustrations related to the objects: item, descriptor or state of descriptor. It is possible to add, visualize or remove one or several images. Furthermore, the manager offers a preview function so that you don't have to check your images before adding them.

Note : It is possible to have two images with the same name. The solution is to rename them before their integration in the KB.

How to open the image manager ?

To open the image manager, click on the button Gestionnaire d'illustrations. The image manager manage the images of the object (item, descriptor, state of the descriptor) selected in the liste.

How to add an image ?

To open an image, select the object in the list, then open the image manager (cf. How to open the image manager?). Click on the button Add an image. Find the image on your hard disk, select it to see a preview on the right zone. You can also add an image stored on the web, click on the button Ajouter une illustration.

Note : It is possible that the image does not appear automatically in the visualization zone. (cf. Why nothing appears when I add an image?)

How to remove an image?

Select an object of the list, then open the image manager (cf. How to open the image manager ?). Select the image in the list, visualize it to check it and click on the button Remove an image.

Note : Only the link between the image and the object will be removed, the image itself won't be physically removed from the folder « images » of your KBd.

How to visualize the images ?

There are two ways :

1- With the image manager :

Select an object in the list, then open the image manager (cf. How to open the image manager ?). Select on of the image in the list to visualize a thumbnails in the right part. Double-click on the element of the list to edit the image in real size.

2- With the "Image" tab of the "Information" zone :

Select an object of the list then right click on the "Image" tab of the « Information » zone. The images will appear as thumbnails if avalaible.
Double-click one of them to visualize them in real size.

What are the supported formats ?

The supported formats are : JPEG, GIF and PNG. We recommand the use the PNG format, which is a free format with a very good quality/size ratio. The images should be in RGB mode and not CMYK.

Why nothing appears when I add an image ?

The image might not appear automatically in the "Images" tab. This is due to a lack of refreshment of the interface. To check the presence of an image, click on another object of the list before coming back to thecright object.

How to create thumbnails ?

It is possible to create all the thumbnails in case of problems (corrupted thumbnail, not updated...). To do so, just go to the « Tools » menu then select "Update the thumbnails".

How does the editor work ?

Editeur
The editor manage definitions of these objects: Item, Descriptor and state of descriptors. Like the image manager, it is possible to browse directly from an object to another with the following buttons :
Previous Go to the previous record.
Next Go to the next record.
First Go to the first record.
Last Go to the last record.

The name of the current object is displayed in the titlebar of the editor.

How to open the editor?

To open the editor, click on the button Define.

How to add a definition to an object?

Select the object to define in the list and open the editor. (cf. How to open the editor ?). Enter the definition in the text zone and click on the button Save to save.

Note : May be a refreshment problem ! (cf. Why nothing appears when I add a definition ?)

How to modify a definition ?

Select the object you want to change the definition and open the editor. (cf. How to open the editor ?). Edit the definition and click on the button Save to save the changes.

Why nothing appears when I add a definition?

The definition might not appear automatically in the "Definition" tab. That might be due to a refreshment problem of the interface. To check if the definition is present, click on another object of the list before coming back to the current object.

Why the loading of my KB take so much time ?

The KB might be long to get loaded. That might be due to the creation of the thumbnails of delocated images (images added via URL). It is possible to desactivate the creation of thumbnail when launching Xper² by modifying directly the file « .preferencexper » located in the folder of the current user (C:\Documents and Settings\Users for windows) : change the line "thumbnailGeneration = true" by "thumbnailGeneration = false".


^Top^

descriptor edition

What is the edition module for?

The descriptor edition module lets you define and enter the descriptive system of a KB:
- Define descriptors,
- Edit the descriptor-states,
- Nest descriptors within descriptors, defining the exclusion criteria,
- Create groups,
- Group the descriptors according to the group they belong to.

What does the interface of Edition of the descriptors look like?

Xper2
The interface is composed of four independant windows:
- The « descriptors » window, used to edit the descriptors.
- The « Groups » window is used to edit the descriptor groups
- The « Etats » window is used to edit the descriptor-states for the selected descriptor on the « descriptors » window
- The « Inaplicability criteria » window is used to edit the conditions of inaplicability for the nested descriptors

When a new KB is created, a descriptor with two descriptor-states is automatically created.

What is the process of edition of the descriptors?

Editing a descriptive system starts with the creation and the edition of descriptors (cf. How to create a new descriptor?) and their states, before nesting hierarchical descriptors (cf. Why structure the descriptors with a hierarchy?) and with the possibly to group them (cf. Why grouping the descriptors?).

How to create a new descriptor?

On the « descriptors » window, right click and choose « New ». Enter the name of the descriptor and save. The newly created descriptor appears on the top of the list and two states « State 1 » and « State 2 » are automatically created. You are invited to modify their name and, if needed, add other states.

If the descriptor is numerical, no states can be added but you can change the measurement unit.

N.B. A descriptor is unique. You cannot give the same name to two different charatcers.

How to modify the name of a descriptor?

Select the descriptor on the  « descriptors » window, right click on it and select « Rename ». Modify the name and validate.

How to remove a descriptor?

Select the descriptor on the « descriptors » window, right click and select « Delete ». A confirmation window will ask you to confirm.

N.B. Trying to remove a parent descriptor will give you the choice between removing only the parent descriptor or deleting the parent descriptor and its children. If you choose to remove only the parent descriptor, the children will be transferred to the next node of the hierarchy.

How to create a new state?

Select the descriptor  to which you want to add a state and right click on the « States » window, then select « New ». Enter the name of the state and confirm. You cannot have two identical descriptors.

How to modify the name of a state?

Select the descriptor-state that you want to modify on the « States » window, right click and select « Rename ». Modify the descriptor-state name and confirm it.

How to remove a state?

Select the descriptor-state that you want to modify on the « States » window, right click and select « Delete ».

A descriptor must have, at least, two states. The upper limit is 63 different states per descriptor.

Why grouping the descriptors?

The rationale of grouping is to be able to assemble a set of descriptors in a common entity. For instance, in a botanical KB it could be interesting to group the descriptors "colour of the bark", "appearence of the bark", "diameter of the trunk" into a group "Trunk".

N.B. A descriptor may belong to several groups.

How to create a new group of descriptors?

Right click on the window « Groups » and select « New ». Enter the name of the group and confirm it. It is not possible two give the same name to different groups.

Attention, in order to right click on the « Groups » window, a descriptor must already have been selected.

How to modify the name of a group?

Select the group, right click and select  « Rename ». Enter the name of the group and confirm it. It is not possible to give the same name to different groups.

How to remove a group?

Select the group, right click and select « Delete ».

How to add a descriptor to a group?

Two possibilities are available in the Edition module:
1- Select the descriptor on the « descriptors » window, then right click and select « Add to a group ». Choose the group from the scrolling list and confirm it.
2- Select the descriptor, then right click on the « Groups » window and select « Add to an existing group ». Choose the group from the scrolling list and confirm it.

N.B. The groups to which the descriptor belongs are displayed in black. Otherwise they are displayed in gray colour.

How to remove a descriptor from a group?

Two possibilities are available in the Edition module:
1- Select the descriptor from the « descriptors », then right click on « Remove from a group ». Choose the group from the scrolling list and confirm it.
2- Select the descriptor, then right click ont he  « Groups » window, select « Remove a descriptor from a group ». Choose the group from the scrolling list and confirm it.

Why structuring the descriptors with a hierarchy?

The notion of hierarchy highlights the fact that other descriptors are applicable only if a descriptor "parent" is described.

N.B. A descriptor can have only one parent. However, a parent can have several children.

N.B : a numerical descriptor can't have a child descriptor.

How to add a child descriptor to another descriptor?

Two possibilities:
1- Select the charatcer you want to mark as a child from the list and drag it to its parent. When the icon of the parent changes ok, you can drop it.
2- Right click on the charatcer you want to mark as parent and select « New ». Enter the name of the descriptor and confirm it. The charater will automatically be added to the list of children.

N.B.: when you select the child descriptor, the descriptor-states of the parent can be previewed in the conditions of inapplicability window. The inaplicability conditions are the states of the parent descriptor that will forbid, if checked, the edition of each particular child descriptor.

(cf. How to edit the conditions of inapplicability?)

How to add a parent descriptor to another descriptor?

Adding a parent descriptor to another descriptor follows the same process as adding a child descriptor. (cf. How to add a parent descriptor to another descriptor?)

N.B.: it is better to declare the parent descriptors first.

How to remove a child descriptor?

The process depend if you would like to:
- Remove a descriptor that is the child of another one. In this case, select the descriptor, right click and choose « Delete ». (cf.)
- Modify the relationship of the child descriptor: select the descriptor and drag and drop it to the right side of the list. The descriptor is moved to the root of the list. When the icon ok changes, you can drop it.

How to remove a parent descriptor?

The process depend if you would like to:
- Remove a descriptor that is the parent of another one. In this case, select the descriptor, right click and choose « Delete ». (cf.)
- Modify the link of the relationship of the parent descriptor in order to release the relationship: in this case release one by one all the child descriptors before moving the descriptor.

How to edit the conditions of inapplicability of the states?

Editing the conditions of inapplicability is to the same thing as choosing the descriptor-states that will make the edition of the child descriptor impossible (for logical reasons in general). Select the child descriptor and check all the states that are displayed on the window « Conditions of inapplicability ».

N.B.: the conditions of inapplicability displayed are those of the parent descriptor!

Is it possible to sort the descriptors?

Yes. Four different sortings are available:
1- Sorting by group Tri par groupe displays a tree-like graph of the groups with their descriptors.
2- Sorting by parent/child dependencyTri hiérarchique displays a tree-like graph all the descriptors.
3- Alphabetical sorting Tri alphabétique of the descriptors.

4- Numerical sorting Tri alphabétique by ascending or descending order.

N.B.: Sorting by group and alphabetically do not allow the edition of the descriptors. To return to the "normal" list you have to sort by parent/child dependency.


^Top^

Edition of Items

What is the item Edition module for?

The item edition module allows:
- Entering the definitions and images of the descriptors
- Entering the definitions and images of the descriptors-states,
- Adding and removing item,
- Entering the comments and images for item

What does the item Edition Module interface look like?

The image below represents the main window of the item edition module:
edition des items

The interface shows three main zones:
- On the left, the list of item
- At the centre, the list of descriptors
- On the right, the list of descriptor-states available for each descriptor and a « Comments » tab.

Each zone contains:
- On the top, a  bar with buttons to open the editor, the image manager, or alphabetically sort the records,
- In the middle, the list of records,
- On the bottom, the « Informations » zone groups four tabs: « Definition », « Images », « Parent/Child » and « Groups » of the selected object.

How to enter the data for a item?

Three steps are required in order to enter a item:
1- Create the item if it does not exist (cf. How to create a new item?)
2- Edition of the item: informations and images (cf. How to add an image?)
3- Edition of the descriptor-states for all the descriptors of the KB (cf. How to edit the descriptor-states of a item ?)

How to create a new item?

In order to add a new item, right click on the « item » region on the left and click on « New ». You will be invited to enter the name of the new item.

Note, Each item as to be unique, hence, it is not possible to have twice a same name.

How to modify the name of a item?

Select the item, right click then left click on « Rename ». You will be invited to enter a new name.

N.B.: Two item cannot have the same name.

How to remove a item?

To remove a item and its description, right click on the item, then left click on « Delete ». Confirm the deletion.

N.B.: The images concerning the removed item will not be automatically deleted.

Is it possible to sort the items?

Yes, it is possible to sort them alphabetically. Click on Alpha Up. Click again on Alpha Down to reverse the sorting.

How to select several item?

There are two possibilities:
- While keeping the « Ctrl » key pressed select the item from the list by left clicking on them.
- Left click on the first item you want to select, then, while keeping the « Shift » key pressed select the last one you want to select. All items between both will be selected

Is it possible to remove several item at once?

Yes, it is. Select several items from the list (cf. How to select several item?), right click and select « Delete ». Confirm for each of the selected items.

How to edit the states of a categorical descriptor for an item ?

Select an item in the list then select a categorical descriptor in the descriptors list. The states of the descriptors are displayed on the right. Deselect the states that are not applicable to the selected item. A descriptor may have several states (therefore, selected).

Note : Edit the parent descriptors first in norder to exclude directly the non applicable children descriptors.

How to edit the values for a numerical descriptor for an item?

Select an item in the list then select a numerical descriptor in the list. Some textboxes will be displayed on the right. Then simply fill out the fields and press the record button (top right) to save them..

Possible values:
- Min : minimum value observed (extreme value)
- Max : maximum value observed (extreme value)
- UmethLower : lower value of the normal distribution
- UmethUpper : higher value of the normal distribution
- Mean : mean
- SD : standard deviation
- Sample : sample size

During identification the range used will be primarily [min, max] if exist, then [UmethLower, UmethUpper] and [Mean - 2SD, Mean + 2SD].

In the case of a unique value (ex : the number of legs) you must enter the same value for the min and max.

Is it possible to simultaneously edit the descriptor-states of several item?

Yes, it is. Select several item from the list (cf. How to select several item?), right click and select the descriptors you want to edit. Check the states that are common to the selected item. This method allows you to  speed up the edition of quite similar item.

Is it possible to compare the descriptor-states of several item?

Yes it is. Select the item from the list (cf. How to select several item?). Then select the descriptors you'd like to compare.
In the « States » zone, on the right, are displayed the selected descriptor-states.
The states displayed in black, either checked or not, are the descriptor-states common to all the items.
The states displayed in grey, either checked or not, are the descriptor-states common to some of the items.

Example :
Etats en commun
Both items have a coarse barke, and not a smooth one. And one of the item has a barke with crevices.

Is it possible to sort the descriptors?

Yes, it is possible to sort them alphabetically. Click on  Alpha Up. Click again on Alpha Down to reverse the sorting.

What is the interest for the tab "Comments" on the right of "States"?

The tab "Comments" is useful to add detail on the duo Item-Descriptor.
For example, for a numerical descriptor, it's important to keep the list of the specimens of the sample.

How to select all the states of a descriptor?

Click on the button Select all on the top right of the button bar.

How to deselect all the states of a descriptor?

Click on the button Deselect all on the top right.

Why can't I add a state for a descriptor?

Apparently, it means that the descriptor cannot be entered. You did not select the right state (inapplicable one) of the parent descriptor that allows you to enter the states of the children. To solve it, go back to the parent descriptor and select a different state, or modify the conditions of inapplicability (descriptors edition mode).

Identification

What is the identification module for?

It allows the interactive identification of an "unknown" item. Interactively means that it is possible to:
- Answer the questions with a choosen order,
- Select several states at the same time,
- Go back in order to modify previous choices,
- Acces the lists of the discarded  and of the still retained item.

What does the identification module interface look like?

Identification
The interface consists of three parts:
- On the left, the list of the descriptors to be described and those already used.
- At the centre, the list of the remainig items and those already discarded.
- On the right, the information zone corresponding to the selected element (Descriptor, Item, and Descriptor-State).

The « Submit » button is used to confirm your choice of the state(s) for the selected descriptor.
The « Reset » button restarts the identification process.

How is an identification carried out?

The identification method proposed by Xper2 is done step by step.

Each step consists of:
1- the choice of a descriptor (either used or not),
2- the selection of one or several descriptor-states
3- the confirmation of the choice by clicking on the « Submit » button.

At each step:
- item are sorted out (as discarded or retained) and the descriptor-states lists are updated, depending on their presence or absence.
- The list of possible descriptors is updated following the choices expressed (display of the child descriptors if an exception state has been chosen).
At any time the remaining or discarded item may be reexamined by clicking on them. On the right, the available images and comments are displayed.

N.B.: When several states are selected, they are considered linked by the inclusicve OR logical operator. If you do not know the anwer to a question, or you have doubts, choose all the states that may be correct.

Note : In case of a numerical descriptor, you can enter the obeserved value in the specific textboxes. All items with the interval [min-max] does not contain the entered value will be eliminated. If the description of the item is incomplete for this descriptor priority will be given to the [Min, Max] values and [UmethLower; UmethUpper] and [mean-2SD, mean 2 SD] to make the range.

What is the « DPower » button for?

The « DPower » button, placed at the top, on the « Descriptors » zone allows sorting the descriptors by their discriminatory power. In order to activate  this sorting, check the check box at the left of the« DPower » button. The descriptors placed at the top of the list are those that discrimniate more item, but are not necessarily the simplest to fulfill !

There are three possible sorting methods :
- XPER original sorting
- Sokal & Michener sorting
- Jaccard sorting

These measures are quantitative assessments of the ability of a descriptor to distinguish between taxa. Here are the formulas and references for these measures. The three discriminatory power result of the sum, for all pairs of taxa, of a measure of dissimilarity between taxa for which the descriptor is evaluated.

For the Xper discriminatory power, the dissimilarity between two taxa i and j for a descriptor is either 0 or 1 depending on whether there are descriptor states in common or not. In the case of discriminatory power based on the Jaccard dissimilarity index or on the Sokal and Michener index, the dissimilarity between two taxa i and i for a handle is between 0 and 1 and depends on the number of states of the commondescriptor.

We can notice that:
n11 states of descriptors present, i.e. possible for both taxa
n10 states of descriptors present for only the first taxon
n01 states of descriptors present for only the second taxon
n00 states of descriptors impossible for both taxa

- the dissimilary according to Xper [Lebbe, 1991] is based on incompatibility between descriptions. Two taxa i and j are incompatibles (or dissimilar or discriminated) if for one descriptor there is no common states of descriptors, i.e. if n11 = 0.
dXper(i,j) = 1 if n11 = 0 otherwise dXper = 0
it means dXper(i,j) = max(0, -n11+1)

- the dissimilarity according to [Jaccard, 1901] take into account at least the states of the descriptors of one of the two taxa i and j that we compare.
dJaccard(i,j) = (n01 + n10) / (n01 + n10 + n11)
Similarity = (1 – Dissimilarity), i.e. two taxa are even more similar than their number of common states of descriptors increase.
SJaccard(i,j) = n11 / (n01 + n10 + n11)

- the dissimilarity according to Sokal & Michener [Sokal & Michener, 1958] between two taxa i and j for one descriptor is measured by taking into account all the possible states existing for the descriptor and not only those existing for at least only one of the taxa.
dSM(i,j) = (n01 + n10) / (n00 + n01 + n10 + n11)
Similarity = (1 – Dissimilarity), two taxa are even more similar if they share common present and absent states.
SSM(i,j) = (n00 + n11) / (n00 + n01 + n10 + n11)

N.B: If the alphabetical sorting is choosen , the discriminatory power sorting is cancelled.

Is it possible to sort the descriptors?

Yes. The classical alphabetical sorting is achieved by clicking on the  Tri alphabétique button, and it is also possible to sort the descriptors according to their discrimatory power. (cf. What is the « DPower » button for?)

How can the the descriptors be filtered by group?

Filtering by group allows focusing on a single group, lightening the list of descriptors to fill. It may also be useful if only an incomplete specimen is available.
In order to filter the descriptors by group, select the group from the list. The descriptors that belong to this group will then be underlined in blue. If the « Filter by group » box is checked, the descriptors that do not belong to the selected group(s) will be removed from the list.


^Top^

Tips and tricks

Save your KB

Save the base in a dedicated folder, a KB contains several files and a folder '/images' which will be saved in the selected location.

Use a filename without any special characters (space, emphasis,...) and for compatibility reasons with the old Xper format, use preferably a file name with a maximum of 8 characters (in MS-DOS, the filenames are troncated after 8 characters).

Keep a backup

Because no one is safe from a hardware problem and because Xper² is not perfect, we recommend to save oftenly the last versions of your base and to store them.

Copy of an item

For a faster edition, if several items are similar (only few differences) you can 'Copy of an item' and make the modifications afterwards.

Note Selecting several states is a logical OR (polymorphism)

Standardize the descriptors and their states!

This step is essential before starting creating a KB! Thinking about the data and the entities related to the studied object will fix the structure of the KB: the groups, the descriptors and most of all the states!

It is even more difficult to re-edit the items after having added and/or modify the descriptors.

Check you base!

Checking your base will tell you if it is correct or not: if all the items are entered, if the items are discriminated and if all the states as well as the descriptors are used.

Descriptors not described: when the descriptors are not entered.
Descriptors described by mistake: when the descriptors are not possible because of the dependency between a parent descriptor and his child. It can happened when you modify relationships between descriptors after having edited the items.

Note This step is essential for the good use of XPER tools (mindescr, makey, distinxp,...)

Use the check base function for a faster edition!

A click on a line helps in selecting directly the values of the edition interface.

Use the HTML tags to define objects

That helps to format the text: bold, italic, underligned, html link,... And might be very useful for web publishing.

Add a comment for a duo (item, descriptor)

To precise the observed value
(ex: for a numerical descriptor, precise the references for the specimens of the sample)

To add furthers information
(ex: precise the uncertaintly of the description, any source,...)

Optimizing the images

The size of the images are very important in term of performance! We recommend to import images below 500 Ko (300 Ko is a good average). The best solution is to resizing the images and/or change the format (a maximal resolution of 1024x768 is enough and gives a good quality/performance ratio).
You can also store your pictures on line (Picassa, Flickr...) and add only the URL (link) to the image.

Beware the files name

For compatibility reasons and web publishing, we recommend to avoid files name with special characters (emphasis, any signs: $, @, ...) and space.

Organizing the 'images/' folder of the base

You can create subfolders in the 'images/' of the base to store your images. This allows to sort and organize the files coressponding to the groups or the descriptors for examples.

Note It is only now that you can link the images to the objects of the base and not the other way round.

Beware the copyrights !

If you are using images from different sources, be careful because they may be copyrighted. You have to ask for authorizations before using them and mainly if you publish them! Some web sites gives free images.

Copyrights on a KB?

You are the only owner of your KB, hence only you can choose the rights on your creations. The aim of a KB is gathering knowledge and sharing it !

Add a comment, a copyright on an image

To add a copyright on the image, you can use a free software that will write a comment directly on the image, very useful for web publishing: a copyright, the author's name, a legend, ...

For Windows: CopyRightLeft is very simple to use!

In case of errors, check the xml format ?

The xml format has a « xpd » version number. The current version of Xper², 2.1, works with the 1.0. Xper² is able to read all former xpd versions. (= descending compatibility)

The xpd file is a xml file taht can be edited with a simple text editor. The structure is as follow :

<xper targetNamespace="http://lis.snv.jussieu.fr/xper/XML xper.xsd" 
	xml="http://www.w3.org/XML/1998/namespace" xpdversion="0.95" 
     xper="http://lis.snv.jussieu.fr/xper/XML/Schema">
<informations>
<firstedition date=""/>
<lastedition date=""/>
<authors/>
</informations>
<base lang="">
<name></name>
<statistics nbind="" nbmod="" nbvar=""/>
<description/>
<variables>
<variable homology="" id="" nbmodes="">
<name></name>
<description/>
<images/>
<modes>
<mode id="">
<name/>
<description/>
<images>
<image url=""/>
</images>
</mode>
</modes>
</variable>
</variables>
<individuals>
<individual id="">
<name></name>
<description></description>
<images>
<image url=""/>
</images>
<description_list>
<desc_element unknown="" var_ref="">
<comment/>
</desc_element>
</description_list>
</individuals>
<groups>
<group id="">
<name/>
<description/>
<groupvar>1/2/</groupvar>
</group>
</groups>
</base>
</xper>

The xpr file is a text file, editable with a text editor. It shows the relationships (in binary code) between the descriptors.

Each block of 3 lines represent one descriptor:
the first is the index of the parent descriptor (0= no parent)
the second line shows the conditions of inapplicability (sum in binary)
the third is the index of the child descriptor

The xpx file is a text file, as editable with a text editor. It shows the relationships (in binary code) between taxa and states of the descriptors.
n lines is the n descriptions (n is the total number of descriptors) for one taxon total number of lines in the xpx file = nb_descriptors x nb_taxa ex : the first line is for the first taxon and the first descriptor, the value is the binary sum of the states corresponding to this duo.

The xpw file is a text file, it contains the weights of each descriptor.

Don't forget the XPER format restrictions

During the export to the old XPER format, you will loose some information because the Xper² format is now richer than it used to be. It allows the management of:
definitions,
images
commentaries on a duo taxon-descriptor
unknown states
numerical descriptors

Besides, the XPER format needs:
14 maximal states for one descriptor.
Strings cannot exceed 60 characters because of DOS restrictions

How to publish a base on-line?

We can host your KB on our server. Just upload your KB HERE.
Otherwise, you can also, create a independant applet that you can put on-line wherever you like it (see Install FAQ).


^Haut de page^

Glossary


KB :
Knowledge Base

FAQ :
Frequently Asked Questions ou Foire Aux Questions

GIF :
Graphic Interchange Format

CAI :
Computer Assisted Identification

JPEG :
Joint Photographic Expert Group

PNG :
Portable Network Graphics

CSV :
Comma Separated Values (http://fr.wikipedia.org/wiki/Comma-separated_values)

SDD :
Structure Descriptive Data (http://wiki.tdwg.org/SDD)
^Top^