I'm building a school MIS, what's next?

[speedypro]

Happy New Year everyone.

 

Thought I'd report on my progress. The website is called Tuicamp. It's still a proof of concept but it's taking shape at the same time. So what's the direction? I'm starting off in a modular fashion: cashless payment, messaging etc:

 

 

1. Meal Planner with a calendar view (for parents)
   
2. Message Board, with commenting
   
3. One-off payments
   

 

Meals are taken from the Easy Lunch Company's 2025 menu.

 

I've made it easy for anyone to sign in. There are two sign ins: school and parent that use the same password on the sign in page. My next step is to approach my child's school to know what they are using, what they hate etc. My first step towards anything. The website is in a "presentable state" which means the UI (how it looks) and function are not final. Things change every week. Feel free to try both log ins, desktop, mobile etc. Only the parent view has dark mode, for now, though this is not important. It's only there because my eyes burn at nights working on this

[Roberto]

Will you be offering single sign-on options? E.g. integration with Google and Microsoft’s authentication services? I’d say that was essential these days, asking anyone to remember another username/password in their lives is unreasonable. Edited January 3, 2025 by Roberto

[speedypro]

Will you be offering single sign-on options? E.g. integration with Google and Microsoft’s authentication services? I’d say that was essential these days, asking anyone to remember another username/password in their lives is unreasonable.

Morning. Yes I will but that's something I need to look into properly. For most parents, they are not tech savvy but could look into something easy for parents/carers. Google and Microsoft’s authentication services are good options.

[Roberto]

For most parents, they are not tech savvy

 

Precisely my point. So reducing the number of usernames and passwords they're expected to remember to interact with their school about little Johnny is a good thing.

 

Doubly so if they have multiple students in the same school - one login to manage little Johnny and little Jessica's lunch in one place is a very good thing. Not just for them but for whoever has to support all this at the school...

[speedypro]

...Doubly so if they have multiple students in the same school - one login to manage little Johnny and little Jessica's lunch in one place is a very good thing. Not just for them but for whoever has to support all this at the school...

 

And that's how it will be

[dhicks]

Yes I will but that's something I need to look into properly.

 

You might want to look at a 3rd-party service to handle Google/Microsoft/Apple/etc SSO. From my own experience, implementing one SSO integration is easy enough, but the more you add the more fiddly / time consuming it gets to both implement and maintain.

 

For teacher / pupil logins, SSO is definatly the way to go. For parents, etc, it can be more tricky. I quite like login-code-via-email - the parent just has to put their email address in, then they get a one-time login code via email. It saves them having to remember / store / reset passwords, and seems to be common enough these days that it should be reasonably familier to people.

[speedypro]

You might want to look at a 3rd-party service to handle Google/Microsoft/Apple/etc SSO. From my own experience, implementing one SSO integration is easy enough, but the more you add the more fiddly / time consuming it gets to both implement and maintain.

 

For teacher / pupil logins, SSO is definatly the way to go. For parents, etc, it can be more tricky. I quite like login-code-via-email - the parent just has to put their email address in, then they get a one-time login code via email. It saves them having to remember / store / reset passwords, and seems to be common enough these days that it should be reasonably familier to people.

It’s a bit tricky with parents as parents are invited/added by the school. I could look at add login with both email and Gmail for parents and Microsoft for school. I could start looking into this next.

[mavhc]

You might want to look at a 3rd-party service to handle Google/Microsoft/Apple/etc SSO. From my own experience, implementing one SSO integration is easy enough, but the more you add the more fiddly / time consuming it gets to both implement and maintain.

 

For teacher / pupil logins, SSO is definatly the way to go. For parents, etc, it can be more tricky. I quite like login-code-via-email - the parent just has to put their email address in, then they get a one-time login code via email. It saves them having to remember / store / reset passwords, and seems to be common enough these days that it should be reasonably familier to people.

 

You think people know their own email address, very funny joke.

 

Having some random person get a login to your account about your children because they don't know their own email address is an interesting GDPR question

[speedypro]

I've looked into the SSO option for parents and it's not possible right now. Remember that the school adds the parent/carer (pc). That pc may have an email address from Apple, Google, Hotmail, any. I wont be going that route to support all those logins. I'll stick with email/password log in, for now. It works perfectly fine with, say, ParentPay. I will be looking into mygovscot and see what's that about.

 

For schools, I could easily add Microsoft. These options would be last on my list.

[dmj]

I've looked into the SSO option for parents and it's not possible right now. Remember that the school adds the parent/carer (pc). That pc may have an email address from Apple, Google, Hotmail, any. I wont be going that route to support all those logins. I'll stick with email/password log in, for now. It works perfectly fine with, say, ParentPay. I will be looking into mygovscot and see what's that about.

 

For schools, I could easily add Microsoft. These options would be last on my list.

 

This really shouldn't be hard. I was under the impression all the major frameworks support social auth. I've certainly set this up before (in Django) and it was a few lines of code.

What backend framework are you using?

[dhicks]

Having some random person get a login to your account about your children because they don't know their own email address is an interesting GDPR question

 

You check the email address entered with your database of parent email addresses before sending an email - I think the project being discussed in this thread is now not neccesarily the main source-of-truth for parent email addresses, so those are probably being synced from the main MIS or a service such as Wonde.

[PotNoodleTech]

Would be beter if you can use not just the parent email address but the parent phone number to confirm ownership. Most systems also need some sort of confirmation of the childs name or date of birth too.

[dhicks]

I've certainly set this up before (in Django) and it was a few lines of code.

 

One issue I've seen before: GMail (and, probably, other email providers) let you associate your own email address (with custom domain) with a GMail address. This means the user can log in with their own email address, but Google's system sees it as a GMail address, and their OAuth system returns the GMail address, not the associated custom address. Trying to explain this to end users is generally difficult - they've often forgotten / not understood that they have a GMail address.

 

In the past, I've found implementing individual OAuth login mechanisms easy enough, but maintaining a set of working OAuth logins is a maintainanace task that does need some attention. Handling just Google / Microsoft logins (possibly Apple?) should cover most school-provided accounts and should be quite maintainable, but trying to cover every OAuth provider around could turn into a bit of a job. There are thinrd-party providers that will handle logins for you, but of course those tend to have a cost associated.

[speedypro]

What backend framework are you using?

 

Using Ruby on Rails. I would still need to support all those accounts ie getting a client ID/Secret to use in my application, and I would like not to do so at this stage. Only Gmail and Microsoft I'm considering.

[dmj]

Using Ruby on Rails. I would still need to support all those accounts ie getting a client ID/Secret to use in my application, and I would like not to do so at this stage. Only Gmail and Microsoft I'm considering.

 

I guess that makes sense, not many parents are going to want github auth

 

Rails is decent, so lots of options there. My only foray into Ruby was with ERB templates in Puppet, so I'm not likely to be much help if/when you do the OSS thing (please do), but I'll happily do the pipelines

[Edu-IT]

With medical conditions / allergens being so important can you add a report to the school side - or some sort of notification - when a parent changes these on a child’s record?

[speedypro]

With medical conditions / allergens being so important can you add a report to the school side - or some sort of notification - when a parent changes these on a child’s record?

 

Glad you brought that up. There's an audit on the student side for when anything change, and by who, it's recorded. There's no UI for it as yet. Good point for the notification. I'd do just that.

[speedypro]

... or some sort of notification - when a parent changes these on a child’s record?

 

 

I'm almost done with the notifications. Notifications will be delivered via email, for now. Hopefully I'll have this live by next week. In a few months I should have have both SSO (MS and Gmail) working with each user with their own demo to play with, privately. At that stage, I'd consider the app to be stable.

[Edu-IT]

Looks good.

[speedypro]

I finally had a call with a Stripe's representative regarding my payment business model. The initial plan was to have a top-up pot for each parent/carer. This "pot" would "feed" a student, regardless of the school they attend, once they use the platform. Stripe does not support the "top-up" feature because there's no where to store that "virtual money". It will be that a parent/carer would need to have a payment source attached: banking card or/maybe a bank account.

 

For children that are not eligible for free meals, their parent will only be liable for payment after a meal was taken. The platform will then attempt to take payment, say, £3.50 for that meal from that parent/carer attached payment source. At this stage, payment could fail. When failed, we'll send a notification to the account holder (parent) of any events until a payment was successful. Yes, parent will still have full control to stop this reattempts if they deem something is wrong but by doing so will open up a dispute with the school and themselves (parent/carer). Any pre-order will become a debit until actioned.

 

I, myself, as a parent would be ok with this once I still have full control of what goes out of my funding source. To recap, there will not be a top-up feature and actual payment will be made once their child has taken a "paid for" meal. One-off items will require upfront payment.

[Rob_D]

I could see it being ok at primary level.

As a lot of primaries have fixed menus published well in advance and often ask parents to pre-order/book their kid's food, I'd rather have the option to pre-pay for the meal ahead of time, ideally alongside booking the meal, with the school/lunch supplier. But post payment isn't a deal breaker by any means.

 

What your proposing probably wouldn't fly at secondary where the more "caffe" style canteen is predominant, as it's basically letting the kid buy as much as they want directly on the parent's credit card/bank account.

 

Could you not have the parents buy credits and then you store their account balance/"virtual money" on your side?

[speedypro]

Now I see the issues. Yes a child could add more items for the parent to pay later which could lead to issues. I'll implement "credits" but this opens me up for chargebacks etc, which I was running away from but I guess it should be ok. In that case, I'll leave the "top-up" feature and find ways to make this work which means the school will then request payment from me. I guess I'd need some Ts & Cs for this method. Good spot. Thanks.

[dmj]

Now I see the issues. Yes a child could add more items for the parent to pay later which could lead to issues. I'll implement "credits" but this opens me up for chargebacks etc, which I was running away from but I guess it should be ok. In that case, I'll leave the "top-up" feature and find ways to make this work which means the school will then request payment from me. I guess I'd need some Ts & Cs for this method. Good spot. Thanks.

 

I think this will be for your benefit: If lots of parents are topping up early in a term, that's a lot of money to invest.

[speedypro]

I think this will be for your benefit: If lots of parents are topping up early in a term, that's a lot of money to invest.

Yes but I'll always be liable for chargebacks, but that should be ok as every seller is liable someway or the other. The top-up is now completed and money transferred to the school works fine, less any fees. Needs polishing before it's available on the website to demo.

[Edu-IT]

How does it work if there’s no credit to charge for a lunch? Will it debit next tike there is a top up?