The question of the system is a matter of saying, and it's a lot of holes. The students' identities, numbers, faculties, dormitories are all at the side of the system, and the billing system has to identify people, tie up rooms, give them privileges according to their identity. But in reality, the system is often built years ago, with old interfaces, typographical distortions, data not necessarily real-time, and hard matching used to be difficult.
Let's get to the bottom of this.
The most easy mistake before the meeting is to pull all the fields in the engineering system. There are few real needs for a billing system: the number of a school or a cartoon is uniquely marked, names checked, the status of a student is tied up in a room without being at school or in a dormitory and at least one identity type separates undergraduate students. The more the field is drawn, the heavier the interface is, the slower synchronous and the greater the probability of error. We suggest that each side confirm the meaning and format of each field and the extra absence of any additional text.
Interfaces are selected according to the actual abilities of the learner
There are no different ways to connect: real-time interfaces call, time batch synchronization, or intermediate library exchange. Which is good, regardless of whether technology is advanced or not, the system can be matched with a learning industry. If a student can provide a stable real-time interface, the authentication will be best; if an education worker has a unstable or unwillingness to open the interface, then it will take a step back and do a full morning increment; if an old system does not have a connection, only an intermediate is used, and both parties agree on a table. We see that schools make a hard real-time interface, which results in a system that cannot handle frequent calls, which is a direct delay in the certification peak period, and finally we change to a fixed time when the sync is steady. Technology programmes need to be realistic and not the other way around.
Who's the number first to die?
The most difficult thing to argue about is who's going to be at the end of a couple. Students have moved their dormitories, and the students have changed their numbers, or the billing fees have not changed, or the other way around? This principle must be fixed before they are connected. Usually identity and accommodation belong to such basic information as a student, while the billing system does not read; but account balances, consumption records and the data generated by these counts remain in place until the school time when the fee system does.
You have to keep up with the student changes.
Students are not static, and every year there are new students leaving school, dropping out of school, and re-entering school, and changing their profession. If the billing system fails to keep up, these changes can be disrupted: graduates’ accounts are still being withheld, fresh students start their studies without a number, and those who have been consummated are tied to old rooms. When designing them, they must consider how to synchronize these changes. It is more stable that, once the academic side changes its status, the bill side is automatically matched by an incremental synchronization or notification of events, such as freezing the account number or suspension of the tuition fee. We suggest that one by one list of common types of changes, confirming synchronization logic, etc., should come up with a temporary solution.
The rest of the logs are to be kept to account.
After the system is docked, when there are student identification failures and miscalculations, it is most feared that both sides will say they are okay. So every data exchange that is done leaves a log: how many times has been synchronized, how many have failed, what have failed. With a log, quick positioning of problems is an issue of engineering data or of costing, without two teams slinging each other. In our projects, where logs are fully performed, average time for trouble-checking can be reduced by half; and missing logs, often with one small problem, look up all day. This input is not very significant and the return is very real.
We'll start with a small scale pilot and then we'll roll it all over.
The risk of two systems being connected is too high. It would be prudent to pilot a building or a faculty, run the identification, room binding, billing deductions, and observe for a week or two, see if data are correct, synchronized and unstable. The problems identified during the pilot phase were small in scope, changed from one to another; the problems discovered after they had been fully rolled out were more dynamic than the whole school. We suggest that the pilot scope should include a university with relatively regular data and a high degree of student collaboration, which could follow through on a building.
The core of the system is not so much technology, but rather a clear description of the fields, modes, master data, changes, logs, and pilots. The quality of the interface between the two systems depends on the level of detail that each team communicates with. The rules are aligned in the previous period, followed by steady operation; the first one is past, followed by endless data fights.