1 |
By now, OpenIsis is mainly used for web publishing of Isis databases. |
2 |
Web developers use the Java, Perl or PHP binding for this. |
3 |
|
4 |
|
5 |
There also is a command line version which serves as a demonstration |
6 |
and test of the various features and can be used as a utility for |
7 |
data import/export and similar tasks. |
8 |
|
9 |
|
10 |
A graphical, standalone user interface similar to UNESCO's WinIsis |
11 |
will be developed during 2003, after the Tcl/Tk binding is finished. |
12 |
See |
13 |
> Status current status |
14 |
|
15 |
* developing OpenIsis |
16 |
|
17 |
In general, there are no plans to reimplement every piece of code ever |
18 |
written for isis. To be of practical value, OpenIsis has to maintain |
19 |
compatibility in the format of the database files anyway. So, one may |
20 |
use winisis or whatever existing import scripts to create and maintain |
21 |
the database, yet deploy OpenIsis' perl interface to run powerful |
22 |
reports and the Java Native Interface to allow queries from a Servlet |
23 |
based web application. |
24 |
|
25 |
|
26 |
OpenIsis will focus on providing tools rather than applications. For example, |
27 |
there will be no attempt to mirror the full functionality of winisis. |
28 |
To achieve this, OpenIsis provides access from the most important |
29 |
programming languages: Java and PHP for the web, Perl for the scripts |
30 |
and Tcl/Tk for platform independent GUIs (partly DONE). |
31 |
All others can, of course, link the lib. |
32 |
|
33 |
* Roadmap for the development of OpenIsis -- as of September 2002 |
34 |
|
35 |
After more than a year of OpenIsis development, |
36 |
experimenting with various options and environments |
37 |
and some feedback, we learned something about what |
38 |
is important to the community and what is not. |
39 |
|
40 |
|
41 |
* findings and goals |
42 |
|
43 |
- A fairly portable C library that can be plugged into Java, Perl, PHP |
44 |
or Tcl proved to be valuable and estimated by several developers. |
45 |
There was not much demand for a pure Java version. |
46 |
- There is high demand for a GUI version similar to WinISIS |
47 |
(cross-plattform, of course) with a safe multi-user capability. |
48 |
- This in turn requires a simple but fast and robust server. |
49 |
Multithreading can greatly improve the throughput. |
50 |
- Users are used to incompatible file formats, formatting and so on. |
51 |
There is little of a standard to follow. |
52 |
So building a next generation system clearly is more |
53 |
important than sticking to the (which?) legacy too much. |
54 |
- The power buried in the similarity between ISIS records |
55 |
and other structures like, for example, E-Mails is not leveraged. |
56 |
- The existing versions of the formatting language are the result |
57 |
of a process of growth, adding feature after feature. |
58 |
The logic is mixed with the presentation (e.g. HTML literals). |
59 |
|
60 |
* conclusions |
61 |
|
62 |
- The multiformat support will be dropped in favour of an efficient |
63 |
support of the "DOS/WinISIS" (packed little endian) structures. |
64 |
- The database kernel has to support basic thread-safety |
65 |
for use in Java and other server environments. |
66 |
- Tcl will be the scripting language of choice: Tcl/Tk for portable GUIs |
67 |
and a limited set of Tcl commands for server side scripting. |
68 |
|
69 |
* flavours |
70 |
|
71 |
- The core library will not contain any particular external technology, |
72 |
so it is as lightweight as possible and can be linked into Java, |
73 |
Perl, PHP, Tcl and the like. |
74 |
- There are binaries in plain, with Tcl and with Tcl/Tk. |
75 |
Any of these can act standalone, as client, as server, or both. |
76 |
|
77 |
|
78 |
volunteers are welcome ! |
79 |
|
80 |
... ToDo |