back to article Sheffield University scales back student system after Oracle integration stumbles

The UK's Sheffield University has abandoned the original design of a £30.4m project to update its student record management system after a vital integration with an Oracle corporate information platform stumbled. The mission to introduce a new Student Lifecycle Project (SLP) dates back to 2014, when the university said it …

  1. Potemkine! Silver badge

    "World class" indeed.

    A board not listening to the IT staff, pouring millions in a project wooed to doom while threatening to fire people because of lack of resources... Board who will put the blame on somebody else and stay unaccountable. Why I've got the feeling to have seen this several zillions times?

    1. hoola Silver badge

      I would not have the confidence to say "listen to IT Staff". It is highly likely that the IT Managlement are very much the initiators of the balls-up and they will not be listening to their techies who actually understand what is possible.

      This is a scenario that is quietly being repeated up and down the country, this happens to have hit the news now but there will be many more in the pipeline.

      Universities are being destroyed as institutions of learning and research by:

      1. Idiotic bean counters

      2. Competitive markets

      3. Blairs "50%" must go to university & fees

      4. Too many places/too few students

      5. Reducing quality (targets, surveys, league tables and more targets (NSS, TES Guardian etc.)

      6. Too many people now in management positions that understand business, not education (see point 1, 2 & 3)

      7 An obsession with "The Student Experience" regardless of whether it actually delivers a useful outcome to the student.

      8 The Research Council, REF and how they throw money about.

      Look at the amount of money that institutions have borrowed to build swanky building on their campus but are cutting staff all over the place.

      Ultimately it is not going to end well as we end up a handful at the very top and a huge pile of mediocrity at the bottom and growing number of mergers or failures.

      1. Terry 6 Silver badge

        I agree, but question whether it's (just) IT staff not listened to. In my experience these projects pretty uniformly fail to actually consult the people who's job it would be to make the bloody system work. Clerks and admins etc.

      2. Alex Stuart

        Don't forget cancel culture also - see recent events https://www.theguardian.com/world/2021/oct/28/sussex-professor-kathleen-stock-resigns-after-transgender-rights-row

        Though the uni did stick by the lecturer here, at least, it doesn't paint a good picture if the protesting/graffiting students went unpunished.

        1. disgruntled yank

          let's be sure to mention that!

          I'm not sure what transgender-rights rows have to do with IT projects, at least not since the days the prudent tech kept a gender-bender in the traveling kit.

          And actually a bit of cancel culture might not hurt the IT world, provided it was losing projects that were canceled, and canceled early enough.

          1. Yet Another Anonymous coward Silver badge

            Re: let's be sure to mention that!

            The Oracle consultant turned up from London, everyone was friendly and called him 'love' - he thought he was in American Werewolf and ran away

      3. Anonymous Coward
        Anonymous Coward

        "Idiotic bean counters" - One University I worked at had a departmental Bean Counter say at a high level finance meeting something to the effect of "I don't care if we produce any research, as long as we break even". It remained in post.

        But the REF is even worse.

      4. Anonymous Coward
        Anonymous Coward

        Let me fix these "facts":

        1. Idiotic bean counters - well, yes, there are financial imperatives that beancounters have to be aware of and the reason for that is that it is the current government policy that HE should be semi-privatised

        2. Competitive markets - Yes, Government policy again.

        3. Blairs "50%" must go to university & fees - yes, not problem in itself, after all who wouldn't want a better education for their children, the issue is how it was implemented in government policy such as introducing fees, removing the student numbers cap and the expectation pf competition in the sector whilst still tightly regulating it so no real competition could emerge.

        4. Too many places/too few students - well, no. Currently there are still more applicants to university across the UK than there are places. There is a separate discussion to be had about readiness for higher education amongst applicants - but that is more to do with secondary education and its focus on exams rather than skills.

        5. Reducing quality (targets, surveys, league tables and more targets (NSS, TES Guardian etc.) - until covid, the league tables have remained largely similar if that is your proxy for quality.

        6. Too many people now in management positions that understand business, not education (see point 1, 2 & 3) - the management burden has increased significantly due to government policy and legislation hence more managers, who are, on the whole more business like. If you are doing a large IT project I think you need business managers not academics and I would suspect that a reason this particular one is TITSUP is because there is too much diversity of process due to a lack of business type management in the university.

        7 An obsession with "The Student Experience" regardless of whether it actually delivers a useful outcome to the student. - Well, Government policy - enter the Office for Students - the explicitly says student experience is paramount but has zero students on the OfS panel.

        8 The Research Council, REF and how they throw money about. - Its pretty hard to get a research grant, the point of research is to do exactly that and sometimes that means failure, and REF provides a proxy for global quality (it has issues though). That doesn't mean the money is wasted when research fails because you have still learnt something, if you want an example of Government waste look at Test and Trace where nothing seems to have been learned.

    2. Doctor Syntax Silver badge

      "World class" indeed.

      Always a harbinger of doom. Why not settle for a working system? It's easier and better.

      1. Terry 6 Silver badge

        Yes, when I read of a new (anything) meant to be "world class" or "world beating" etc. I automatically assume that it will start by ignoring all the existing experience and try to create something from the ground up, totally different to the systems that at least are known to work.

        1. Anonymous Coward
          Anonymous Coward

          < Yes, when I read of a new (anything) meant to be "world class" or "world beating" etc. I automatically assume that it will start by ignoring all the existing experience and try to create something from the ground up, totally different to the systems that at least are known to work. >

          1000 upvotes.

          Any university has the same IT needs as a hundred-odd other universities around the country. Procurement should be gathering the experience elsewhere, seeing which is working the best, and saying "we'll have one of those - just change the logo on the login page". Nothing experimental, world-beating, customised etc .

          The same principle goes for local authorities, police forces etc whose IT system upgrade travails are reported so often in The Register.

          1. Terry 6 Silver badge

            Thank you. IMAO instead of world beating there needs be just a short set of requirements;

            *What's the job; as defined by the people actually getting it done, and not some bloody executive sitting in an office, or Heaven Forfend an outside consultant who has even less of a clue

            *What can IT do to make sure the job gets done efficiently (as opposed,arguably, to writing everything on a few sheets of A4)

            *What expansion or change might we need to allow for over the next 10 years of so

            *What are the maintenance and development costs over those 10 years

            I know the last one will always be an underestimate - but even that's better than no estimate ( read, budget) at all, which seems to be the usual way of doing stuff, in any project not just IT. I call this the "School Blinds Issue". For a reason. In the mid 80s I used to do some work in a school designed (50s I think) with South facing windows and some clever external blinds that made the rooms usable when the sun shone. I've since come across a few of these from that period btw. These needed to be maintained, they had thick strings ( like a fine rope)and pulleys. By the time I visited these blinds hadn't functioned within anyone's memory. They'd never,as far as anyone could tell had the ropes replaced (many of them had long since become brittle and snapped) let alone the pulley wheels - most of which were totally frozen. As far as anyone could ascertain there'd never been a budget allocated for this maintenance by the LEA, not even a schedule or a plan. They'd just built the damn thing and forgot about it..

        2. Doctor Syntax Silver badge

          The one thing that is world-beating is the level of Dunning-Kruger over-confidence that drives that approach.

        3. Anonymous Coward
          Anonymous Coward

          ...when I read of a new (anything) meant to be "world class" or "world beating" etc. I automatically assume that it will start by ignoring all the existing experience...

          <p>

          And almost universally, the drivers of those projects are of the technical aptitude and experience to say that programming problems are "simple" and to argue that you should have something done "in a couple of weeks" when you tell them it will take 3-6 months, once staff can be scheduled.

      2. WhiteDragon43

        World Class Mediocracy

        Blighted by our Born to Lead Old Etonians and other bum boys from the Bullingdon Club. Nothing equitable at all in the race to the "bottom" - a wonderful achievement to encumber £9k a year debt to many for a certificate of mediaocracy.

    3. Sharik

      It's also unlikely that, once integrated, the system would actually improve the administration of student records. Every time I've seen a new system introduced a poor user interface has meant that it takes longer to perform normal tasks than it did with the old system. My particular favourite was the shiny new system that doubled the number of actions required to input each individual student assignment mark.

  2. A Non e-mouse Silver badge

    Sheffield University and College Union (UCU) said the SLP project had spiralled out of control with seemingly little governance oversight

    Glad to see it's not just the Oxbridge elite that can screw up enterprise systems.

    1. This post has been deleted by its author

    2. Anonymous Coward
      Anonymous Coward

      >Glad to see it's not just the Oxbridge elite that can screw up enterprise systems.

      I was working at Caltech when they introduced a new Oracle system

      It was so bad we lost a Nasa mission because we were so late in paying some vital sub-contractors that they went bust and the project eventually got cancelled.

      We ended up having to pay cash for liquid nitrogen fills from our own stores.

  3. Anonymous Coward
    Anonymous Coward

    Just get ready

    for the inevitable visit by the Oracle Audit Goon Squad.

    Be prepared to cough up a load more money for bringing their wonderful software into disrepute. (sic)

  4. Yet Another Anonymous coward Silver badge

    How much?

    Unless Sheffield has got a lot bigger it can't have more than 30,000 students - so that's a grand per database entry.

    They are aware you don't need a separate computer for each record?

  5. Bitsminer Silver badge

    30%

    30 per cent of the scope will remain undelivered, this does not pose a major risk as many of the requirements are no longer relevant or viable...

    Changing or deleting 30% of your requirements in 5 years is mind-bogglingly....little. I would have expected 1% per month, or 60% after five years. And a competent system management team with a budget to match that rate of change. But, you know, "magic" is defined as anything beyond your understanding.

    I suspect the problem is really that education, like medicine, sees IT as an asset to be purchased once and thereafter neglected, like a building. Instead, IT is a liability that needs to be continuously supplied with attention, care, adaptation, nurturing and money. Lots of money. Like a Hollywood starlet.

    1. lglethal Silver badge
      Trollface

      Re: 30%

      Saying IT is like a Hollywood Starlet! That's a bit harsh....

      ... on the Starlet.

  6. devin3782
    FAIL

    Of course it failed Oracle was involved, they don't need it to succeed, just for you to pay their licence fees.

    £30 million, what a colossal waste of money £30M could have given under paid staff a raise! Oh but of course, different budgets (slush funds for mates)

    I'm sure I could have built a system for student records with another DB tool for way less than £1million, University SQL modules usually show examples of a student database system anyway surely any developer who's been to uni could model this data with their eyes closed after quick check of their course notes.

    1. Yet Another Anonymous coward Silver badge

      Or they could have just copied a system from another university.

      What if there was a way of creating free software collaboratively and licensing it in such a way that lots of users could take it and adapt it for their own use and contribute changes back?

      I suppose the student processing requirements for Sheffield university are so vastly different from those of Sheffield Hallam university that only Oracle could handle the technology transfer

      1. Ken Hagan Gold badge

        It's a fascinating idea, but it could never survive in an academic environment.

  7. Anonymous Coward
    Anonymous Coward

    Raising the Titanic would probably be easier.

    I work in IT at a university, and I know pretty much how this gets to this point.

    Russel Group unis like Sheffield are mostly run by academics and they generally have no idea how processes work. They all think they know best and that their individually crafted approaches to each individual problem is the right way to do. In addition, they know how to forcefully argue their case, because that is the job of academic. Consequently, what you should never do when doing an IT project in a university is ask them what their requirements are because what you will get is 5,000 different ways that you need support a process and the expectation is that they will all be supported because the academy (and UCU) has nearly all the power in a university. Any big ERP like Oracle E-Biz suite is just not needed: its too configurable, its too burdensome and its too expensive. Tribal SITS is a widely used student records system but it has its own problems, not least it is written in a 4GL from the 90s that has been updated to support web pages and it allows extreme configuration making almost every implementation at every university different. Consequently, there is no simple integration route with an on premise SITS; you are either using tight integration with database links or you need a separate ESB layer somewhere to mediate the integrations. Oracle e-Biz suite doesn't play nice with anything that isn't using Oracle tooling. SITS doesn't.

    The right approach to systems implementation in universities is to have standard best practice processes and tell the business to change and modernise. That way the processes are simpler, the behaviours area predictable and nearly all universities have exactly the same requirements for back office systems which is to accurately report student attainment, student and staff numbers, and financial records to the regulators in government. Its the same requirement for every uni in England (other countries are different but also uniform). Trouble with that is that modernising a universities working practices is like raising the Titanic, not impossible but extremely difficult.

    1. Yet Another Anonymous coward Silver badge

      Re: Raising the Titanic would probably be easier.

      > have standard best practice processes

      We had that, "standard best practice" was that all the IT infrastructure was to come from a single supplier. Naturally the only supplier that could offer mainframe + mini + workstation + PC (back in the 80-90s) was IBM.

      So we had a single "integrated system" that had an IBM mainframe running God-knows-what, AS400s for HR and accounting, RS6000 unix workstations and PS/2s for student desktops.

      Strangely this didn't all work together seamlessly and was gradually abandoned. But only after getting rid of all the sock+sandal wearing beardies in this basement that knew how everything worked before.

      1. Peter Gathercole Silver badge

        Re: Raising the Titanic would probably be easier.

        Was this the Bloomsbury Computing Consortium serving various London colleges, by any chance?

        I was the IBM Support Advocate in the IBM AIX Systems Support Centre for that customer for a few years in the early '90s, and they had many problems with IBM integration, because there just wasn't really any between the different platforms, even those running AIX.

        They had AIX/370 on a mainframe, AIX 3 on RS/6000 and AIX 1.2 on PS/2s. And the teaching problems they had were many.

        AIX/370 and AIX/PS2 had a clever feature called Distributed Computing (AIX/RT on the 6150 also had it), which allowed you to have shared network filesystems (not NFS or RFS) and use the same PATH for the different architectures, and the exec system call would pick the correct binary for the architecture running (AFS had a similar feature). But AIX on the RS/6000 didn't, which rather sabotaged that clever integration.

        In addition, BCC wanted to use both RS/6000 and PS/2 for language teaching, but IBM being IBM had two different strands for compilers, XL and VS. So, the RS/6000 had XL Pascal, and the PS/2s had VS Pascal. XL Pascal complied with the ISO Standard, and VS to the ANSI standard.

        There were a couple of important differences in the standards, although memory fades a bit, I believe one of them was around variant records.

        I had problem records for both compilers sitting on my queue, essentially saying that Bloomsbury wanted the same behavior on both compilers. Quite reasonable, really. They didn't care which behavior (although they would have preferred the ISO one), they just wanted them to be consistent.

        I battled for many months with the L2 team for AIX/PS2 and the L3 team for AIX/6000 (we provided Level 1 support for AIX/PS2 and Level 2 support for the RS/6000). Both teams said that their implementation was 'working to the standard', which they were. But it didn't help the customer.

        I moved off the customer support queues, and dropped the advocacy role before this was resolved. I think the person who picked them up just told the customer that there was no solution and closed the calls. Great customer service!

        But I can't have had too bad a reputation with BCC. A number of years later, I was offered a job by one of the people who used to work there and had moved on, just on reputation and without an interview. They just wanted to confirm I was the person they had dealt with, and asked (within reason) what package I wanted. The recruitment agent was gobsmacked! And I did not take the offer, because I did not want to work in central London.

      2. Anonymous Coward
        Anonymous Coward

        Re: Raising the Titanic would probably be easier.

        > We had that, "standard best practice" was that all the IT infrastructure was to come from a single supplier. Naturally the only supplier that could offer mainframe + mini + workstation + PC (back in the 80-90s) was IBM.

        Im talking about best practice processes, not kit. Standardisation of kit is different issue and should be focused around support needs in my opinion.

        1. Peter Gathercole Silver badge

          Re: Raising the Titanic would probably be easier.

          Sorry, I was just looking for an excuse to recount a story. Pure narcissism on my part.

    2. thondwe

      Re: Raising the Titanic would probably be easier.

      Generally "off the shelf" Student Records Systems are niche and there's not huge bucks to be made - hence the company's involved are relatively small. SITS is well known, but is a bag of nails. Most Uni's customize it and then get tied in for years as it's a non-trivial job to switch horses.

  8. ecofeco Silver badge

    Oracle?

    I think I see the problem.

  9. Anonymous Coward
    Anonymous Coward

    Final Year Project

    Students could probably knock one out. Before their enthusiasm is dulled by years of drudgery.

    1. A K Stiles
      Joke

      Re: Final Year Project

      I suspect many do, on a regular basis, but how is that going to help with a database system?

      1. Korev Silver badge
        Coffee/keyboard

        Re: Final Year Project

        Love it!

      2. Anonymous Coward
        Anonymous Coward

        Re: Final Year Project

        I line 'em up ... you knock 'em ...

        ... in.

  10. smot

    Oracle?

    For 30,000 students? Government departments made do with Excel for reporting Covid results. Surely they can share that success and save money.

    1. Anonymous Coward
      Anonymous Coward

      Re: Oracle?

      Those government talking heads are just trying to throw up pretty pictures on the screen that are too small and on the screen for too little time for the audience to see them. If you feel like downloading and viewing them, well, there just might be the occasional compatibility issue... if the downloads are working today. It depends on the phase of the moon, the direction of the wind, and how many dogs you can hear barking.

      Key point is: never assume a spreadsheet is accurate. Ever. Ever. Ever.

      It is MUCH safer to assume they were knocked out in an afternoon or an evening and have at least three or four glaring bugs in their calculations. And that is presuming the data fed to the sheets was valid in the first place. Spreadsheets aren't notorious for integrity checking their inputs...

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like