AUK Finances and Website Project was: AUK Chairman Statement

[Greenbank]

More generally, the easiest way to work out what a complicated computer program actually needs to do is by writing the complicated computer program and then trying to use it.

And the first version to be written should always considered the prototype. That gives you the knowledge to write the proper version next time round.

Giving a prototype to a completely new development team means they get to write their own prototype (rather than a polished version).

[quixoticgeek]

More generally, the easiest way to work out what a complicated computer program actually needs to do is by writing the complicated computer program and then trying to use it.

And the first version to be written should always considered the prototype. That gives you the knowledge to write the proper version next time round.

Giving a prototype to a completely new development team means they get to write their own prototype (rather than a polished version).

The number of times prototypes become the live version is rather shocking.

J

[psyclist]

Surely it can't be too intensive a job to have a module that crunches a member's finished/not finished data to render awards.

Fitbit does it brilliantly.
Every time you pass a personal milestone a snazzy virtual medal pops up on your phone, in real time.  I'd love AUK to emulate that.

That would be rather cool. Given the appropriate data architecture, quite conceivable to have a well done message (subject to human validation where necessary) on completing an RRTY, an SR, or any award really.

[bludger]

That's the kind of thing I've been alluding when I say there's a huge amount of backend stuff that the current site does that only organisers see.

All of this was done is various manual ways before and centralising it on the AUK website was seen as a huge productivity boost to the organisers and events secretary. I don't think any of those people want to go back to the old ways of doing it.

I will literally eat my bike if a Microsoft website package wouln't already have that functionality. See https://appsource.microsoft.com/en-us/product/dynamics-365/intelligentsystems.eventmanagement-05042019

For real I will livestream me angle grinding my 853 frame into dust, cook it into a cheese sauce, and eat it with spaghetti.
 
They've had risk management features in their stuff for ages https://support.office.com/en-gb/article/add-a-risk-to-a-project-in-project-online-7aa1acc9-50cf-4f15-ac3b-fedf41b31c83

[Greenbank]

That's the kind of thing I've been alluding when I say there's a huge amount of backend stuff that the current site does that only organisers see.

All of this was done is various manual ways before and centralising it on the AUK website was seen as a huge productivity boost to the organisers and events secretary. I don't think any of those people want to go back to the old ways of doing it.

I will literally eat my bike if a Microsoft website package wouln't already have that functionality. See https://appsource.microsoft.com/en-us/product/dynamics-365/intelligentsystems.eventmanagement-05042019

For real I will livestream me angle grinding my 853 frame into dust, cook it into a cheese sauce, and eat it with spaghetti.
 
They've had risk management features in their stuff for ages https://support.office.com/en-gb/article/add-a-risk-to-a-project-in-project-online-7aa1acc9-50cf-4f15-ac3b-fedf41b31c83

I'm not sure what point you're trying to make with that.

Do Microsoft have an AUK compatible events planner as a drop in replacement? No, of course not.

Could the Microsoft stuff you've linked to be configured/developed to do what is necessary? Yes, I'm sure it could. But what's the difference between engaging Microsoft to do that, and engaging F1/IIP to build a bespoke website? Both involve paying someone to get something to do what you want and a suitable support/maintenance contract.

[FifeingEejit]

The problem with a prototype is you first have to understand the business needs to produce it.

Having a previous version of the software as prototype encourages recreating the same thing again with many of its known flaws. Which is fine if you plan a direct rewrite and target tidying bits up in the process.

Prototypes that become the real system, yes...............


Sent from my BKL-L09 using Tapatalk

[bludger]

Do Microsoft have an AUK compatible events planner as a drop in replacement? No, of course not.

Well no because we haven't contracted them to prepare such a thing. But if we did they could do it and I expect it wouldn't cost £200,000 because it would be a matter of combining pre-developed modules, not someone staring at a blank sheet of paper. I expect an AUK website built by MS would involve scarce little actual manual programming.

what's the difference between engaging Microsoft to do that, and engaging F1/IIP to build a bespoke website? Both involve paying someone to get something to do what you want and a suitable support/maintenance contract.

1. Microsoft already the modules developed with which to produce such a website, as they have done for thousands and thousands of clients all around the world.

2. They dont go into administration leaving the client in the lurch!

[Greenbank]

The problem with a prototype is you first have to understand the business needs to produce it.

Not strictly necessary when you start to build the prototype (but obviously useful if you do), but the point is you should understand the business needs by the time you've finished it. If you don't you, almost certainly, haven't finished it properly.

Having a previous version of the software as prototype encourages recreating the same thing again with many of its known flaws. Which is fine if you plan a direct rewrite and target tidying bits up in the process.

Only if you choose to. We either fix the flaws in the prototypes we make here or, where they're structural enough that fixing them would be too great an effort we just make sure we don't repeat the same things in the rewrite. Writing the prototype in the same language as the end system also gives lazy developers a chance to just copy-and-paste much of the existing code - we tend to write our prototypes in something we can't ship specifically to avoid this.

Anyway, this is verging even more off topic.

[Martin]

I don't know if anyone has mentioned the events planner yet.  Each event has an organiser multi-page interface into which they add controls, risk assessments, and all the public details such as cost, facilities and descriptive blurb. The calendar is extracted from the planner. Then finish lists have to be checked, completed and sent for validation. 

None of this has been touched at all by the new project.

yes it's an incredibly useful and fairly complex bespoke bit of kit without which organisers would be back to pen and paper or Excel at best. Not visible by non-organisers but essentially when anyone looks at the finish list,, complete with the name of the event the AAA even the start location this has all been generated in Event Planner. Also it can be copied across from one year to the next all that needs changing are the dates and any info controls if it's the same route.

IIRC it's not even held on the "old" website but the one before that. IIRC it also used to all be done by one man?


As someone who's currently going through a painful IT procurement for a new system that doesn't work as well as the existing one I'd have to say I'd be tempted to keep the old website running by whatever means,  if it can't be reasonably migrated over to the new site

it ain't broke (yet!)

There must be loads of old systems still running on obsolete platforms, for example every time I check in at an airport they scan the boarding pass and it flashes up my name on what looks like a dumb terminal; they even print out the passenger manifest on a dot matrix printer!

[Greenbank]

Do Microsoft have an AUK compatible events planner as a drop in replacement? No, of course not.

Well no because we haven't contracted them to prepare such a thing. But if we did they could do it and I expect it wouldn't cost £200,000 because it would be a matter of combining pre-developed modules, not someone staring at a blank sheet of paper.

F1/IIP weren't staring at a blank sheet of paper. They've developed websites before and they're using pre-developed modules (Umbraco CMS, etc). You're making a big assumption about the cost from MS too, also remember that F1/IIP originally quoted £120k or so.

I expect an AUK website built by MS would involve scarce little actual manual programming.

Then, to put it bluntly, I don't think you understand the requirements of the AUK website, and there's little point continuing this discussion.

[bludger]

Yes what would I a lowly auk member have to contribute concerning the future of an it project I've been paying in for that is on course to outsize the cost of a house 

If the project is untenable then it is time to massively draw down the 'requirements' of AUK and reduce the aspirations of this plan.

[Greenbank]

Yes what would I a lowly auk member have to contribute concerning the future of an it project that is on course to outsize the cost of a house :rolleyes:

What's your suggestion then? So far it looks like it is to see if MSFT might be able to magically do it all for cheap? What if they come back to AUK and say £200k, meaning the entire project will have cost £400k? Or do you expect they'll come back and say it'll be under £50k?

£200k is already spent and gone, for not very much in return, there's no getting it back.

For the record, I think my suggestion is quite clear:-
a) Using the collapse of F1/IIP as a point to stop throwing money at things and reassess.
b) Use the existing volunteer base[1] to shore up the existing aukweb system without any grand rewrites (just do enough to get php/mysql/OS/etc up to the latest levels)
c) Use the existing volunteer base[1] to come up with an assessment of how much time/effort is involved in enhancing the existing system in situ to get it to a point where a new front end could be put on it by an outsourced company
d) Get some quotes to see how much it would cost for a web development company to redo the front end without needing to touch much of the back end - hopefully they can use some of what F1/IIP have done already as AUK has secured the IP for this.
e) Then make a decision on whether to do some/all of the work unearthed in (c) and whether to spend more money or not on (d).

1. By this I mean the people that volunteered to help with this when this all kicked off a few years ago, I don't just mean FF and the existing people who have access.

[bludger]

Yeah I'd talk to a customer success senior at MS. Offer some ad space in arrivée and on the website to partly offset the cost. Try to kid them that some other audax clubs around the world might then take them on as service providers since we're one of the biggest and ugliest (if not the biggest, I can't remember).

If it comes down to it yeah I'd pay £200k, since 200K has got us 'this far' then another lot would represent a massive improvement in VFM. Another 200k per 8000 members is £25, factor in a gift-aidable donation drive and it's achievable. Why not. We'll do it live.

[Jaded]

What might the monthly M$ charge be? The basic default is £100 per month.

[bludger]

Split that between 8000 members that's 15 p per member per month.

[Jaded]

Split that between 8000 members that's 15 p per member per month.

So what - doesn't get anywhere near answering the question. Plus it is 15p a year, I think...

Why not just say "I'm not sure", or "I don't know"?

[bludger]

Well I don't know, but if your guess of 'the default' is £100/month then the sensible thing to do is to contextualise the guesses with the cost implication to the individual paying club member over the course of the month. This isn't even factoring in the non-member fee paid by those signing up for rides without membership. If it was double, triple, or quintiple your estimate of the default then it's still pretty small potatoes for the member paying £18 in annual fees.

[Ben T]

parkrun is probably a much simpler model. Person X did run Y on date Z in time T is the bulk of it (you may think AUK's model is roughly the same but it really isn't)

Maybe that's how it should be.

I personally think this is a good point to be honest - AUK's business logic is currently fairly complicated, but does it need to be?

Why is it actually necessary to have an automated system where

Rider has achieved award X if they've done this set of rides, of this sub-type or this sub-type but not this sub-type, within these dates, except these rides that are tagged like this, oh and one of the rides that could qualify could actually be made up of 2 different rides (a calendar ride and an ECE) so you need to take that into account.
Oh, and some awards have sliding time windows (i.e. a B25000 or whatever it is that has to be done over a 6 year period) so you need to check each possible sub group of rides to see if any subgroup qualify for the award, etc.
Oh, and for this award (RTTY) this ride can definitely count for this month or possibly be counted for the following month but it needs a human to confirm that at least 200km of the ride was done in the later month, etc, etc.

What is the cost benefit analysis of business rules such as the above - i.e. if it costs £x to develop/maintain business rules such as the above, how many riders (y) would leave if you didn't have it?
If £x > y x £membership fee, then it isn't worth it.

The big step (that for whatever reason hasn't been taken) is breaking the perception that what it must continue to do is defined by what it currently does, i.e. that all functionality must be kept.

If you just said - right, from now on, in the new website, we are just going to have: (to paraphrase) Person X did ride Y on date Z. And that's it.
How many people would actually say right, well that's it - if I can't have award X for having done 2 different rides one of which was in a certain time window then it's just not the same to me, so I'm upping sticks and leaving.

You could even simply use smoke and mirrors:
An "award claim service" which simply has a multi-line text box asking riders to list the rides they've done and a big submit button at the bottom to claim the award.
Oh but pray do tell, what incredible algorithm could possibly parse this text to decide whether the rider is worthy of the award?

Code: [Select]

function DoesRiderGetAward(string claimText)
{
  return true;
}
No! But riders would simply claim anything and everything!
Would they though? Really? Aren't we basically taking their word for it anyway that they've even done the ride at all in the first place?
There's no-one checking.

You simply display the certificate and a "well done" message, send the badge if they've ordered/paid for it, but just display a caveat that "claims will be reviewed on a regular basis for integrity".
This "reviewing" can then either be filled by as much volunteer time as anyone who can be arsed wants to put into it, or simply serve as similar to the tv license detector van.

You can still let organisers write complicated rules for the criteria for an award, but why do you need an automated algorithm to decide whether people have fulfilled it or not? Just allow claims and (claim to) audit a percentage of them manually.


This might seem a bit like making it all too simplistic but you aren't actually taking anything away from the member experience.
You could claim an award before, you still can now.

I'm slightly with S2L on this that there is no inherent value in that particular code.
To say there is value in the actual code, rather than simply the maintenance of the awards, requires believing in the notion that members would actually leave AUK if it doesn't have an algorithm that automatically rejects invalid claims. Taking this to its logical extreme implies they would put it to the test and base their membership renewal decision on the effectiveness of the algorithm, which when you think about it is a bit silly.
This might seem a bit of an arcane way of thinking about it, but it's really nothing more than a creative way to avoid cost.

[bludger]

You could even simply use smoke and mirrors:
An "award claim service" which simply has a multi-line text box asking riders to list the rides they've done and a big submit button at the bottom to claim the award.
Oh but pray do tell, what incredible algorithm could possibly parse this text to decide whether the rider is worthy of the award?

This is exactly how they do it in Australian audax. You want an award, well strewth mate fill in the form and email it in to the bludger who checks the inbox a few times a year while sipping a VB, wearing their akubra hat with the corks dangling off it. They offer a refreshingly streamlined series of awards. https://www.audax.org.au/public/audaxawards

I've personally done clerical freelance work, stuff just like this, for £12 an hour. Crunching through spreadsheets on Google drive with my funk jazz playing in the background.

[Davef]

Yes what would I a lowly auk member have to contribute concerning the future of an it project that is on course to outsize the cost of a house :rolleyes:

What's your suggestion then? So far it looks like it is to see if MSFT might be able to magically do it all for cheap? What if they come back to AUK and say £200k, meaning the entire project will have cost £400k? Or do you expect they'll come back and say it'll be under £50k?

£200k is already spent and gone, for not very much in return, there's no getting it back.

For the record, I think my suggestion is quite clear:-
a) Using the collapse of F1/IIP as a point to stop throwing money at things and reassess.
b) Use the existing volunteer base[1] to shore up the existing aukweb system without any grand rewrites (just do enough to get php/mysql/OS/etc up to the latest levels)
c) Use the existing volunteer base[1] to come up with an assessment of how much time/effort is involved in enhancing the existing system in situ to get it to a point where a new front end could be put on it by an outsourced company
d) Get some quotes to see how much it would cost for a web development company to redo the front end without needing to touch much of the back end - hopefully they can use some of what F1/IIP have done already as AUK has secured the IP for this.
e) Then make a decision on whether to do some/all of the work unearthed in (c) and whether to spend more money or not on (d).

1. By this I mean the people that volunteered to help with this when this all kicked off a few years ago, I don't just mean FF and the existing people who have access.

I would agree with this as a strategy and happy to help. I own a few successful websites.


Sent from my iPad using Tapatalk

[Pickled Onion]

Perhaps some kind of new monastic order focused on IT work, where instead of brewing beer or keeping bees the monks and nuns would build and operate websites?

You mean a scary devil monastery perhaps?

I recovered, I haven't posted there for over 20 years.

Me too, I still look in occasionally to see how Chromatic is doing.

[FifeingEejit]

You could even simply use smoke and mirrors:
An "award claim service" which simply has a multi-line text box asking riders to list the rides they've done and a big submit button at the bottom to claim the award.
Oh but pray do tell, what incredible algorithm could possibly parse this text to decide whether the rider is worthy of the award?

This is exactly how they do it in Australian audax. You want an award, well strewth mate fill in the form and email it in to the bludger who checks the inbox a few times a year while sipping a VB, wearing their akubra hat with the corks dangling off it. They offer a refreshingly streamlined series of awards. https://www.audax.org.au/public/audaxawards

I've personally done clerical freelance work, stuff just like this, for £12 an hour. Crunching through spreadsheets on Google drive with my funk jazz playing in the background.

So for comparison, how many members does Audax Oz have and how many SR claims were there last year?

Sent from my BKL-L09 using Tapatalk

[Davef]

Well I don't know, but if your guess of 'the default' is £100/month then the sensible thing to do is to contextualise the guesses with the cost implication to the individual paying club member over the course of the month. This isn't even factoring in the non-member fee paid by those signing up for rides without membership. If it was double, triple, or quintiple your estimate of the default then it's still pretty small potatoes for the member paying £18 in annual fees.

I don’t suppose you have link to where Microsoft offer custom website development ? Are you referring to the template websites produced by wix.com or something else ?


Sent from my iPad using Tapatalk

[bludger]

1200 members as per 2018. So about 1/6 the size.

I don't know about SR claimants but I don't see much reason for their proportion of SRs to differ from ours. So for us we have about 750 a year in a PBP year and about 450 in a normal year. I expect their SR stats are more consistent as they're further from France, so probably about 75 a year?

[bludger]

I don’t suppose you have link to where Microsoft offer custom website development ?

https://www.microsoft.com/en-us/industry/services