Accreditation is a big pain in the ass, as we all know. I agree with many of the other thoughts. Here's my calculations. I have no sentimental attachment to being accredited and if we could offer a good program without accreditation then we should.
- Accreditation forces us to have stuff in our curriculum that we don't want. There is no natural champion for, in particular, the NSEs.
- Students have a strong identification with "engineering undergraduate culture" (e.g. EngSoc) even if many students do end up switching to CS especially after first year. Being iron ring eligible is important to some subset of our students.
- SE has a distinct identity from CS, in a much stronger way than the SE option. To expand on what I said above, I think the principal drivers of that are the cohort system and the iron ring.
- I don't think actual curriculum content is that important to students, and I think it doesn't strongly affect what graduates are able to do 5 years down the road. My assumption is that a graduate can learn what they need to in terms of technical skills. Heresy alert: I'm skeptical that SE 463/464/465 contribute significantly to our students designing and writing better systems later on. On the other hand, the mandatory 3-course FYDP sequence does, I think, cause many students to stretch themselves in ways that they wouldn't otherwise.
We do talk about ethics perhaps a bit more than CS students get, but I also wouldn't put a lot of money on graduates internalizing engineers' responsibility to society. Well, I'd put $1000 on it, but I wouldn't put half a year's salary on it.
OK, so what if we had a cohorted program which had no iron ring? What would I tell people at OUF? Well, I could say the words about SE 463/464/465, but I wouldn't. I could say words about ECE 124 and ECE 222 and getting a bit more Computer Engineering education. We're moving away from them getting as much technical depth as 1st year computer engineers in hardware topics, so that's fine, but the words wouldn't be quite as accurate. I think the words I would believe most would be about FYDP and about the cohort system. It would definitely be more of a challenge to be convincing. We probably would still get enough students.
I think, by the way, that students would duck SE380, but I don't think the most academically inclined top 75% would duck CS 343. Sometimes I'm wrong about these things.
Derek, you've said previously that CS + DHW is a better SE degree than CS. We could move that way if we weren't accredited.
The last question is whether we'd need to remove "engineering" from the name.
To sum up, though, I kind of have a "why change it if it's not broken" attitude towards this question. We could probably make an unaccredited program work, but it's not at all clear that it would be long-term viable, and it's also not clear that it would help students, except that we can remove some NSEs. And we'd lose some things that are important to some students.
[I still want to write the SE retrospective, but writing papers and grants that are in my core research area keep on getting in the way; hopefully next month.]
pat
On 2022-10-17 04:27, Charles Clarke wrote:
There's also the very practical consideration that without accreditation (i.e., the need for SE to be partially supported by an Engineering program) SE will likely melt away into CS. Mary seems to like the SE program, which continues to attract good students. If she wants it to continue to be an Engineering program, it needs to actually be an engineering program.
From: Joanne Atlee [email protected] Sent: Sunday, October 16, 2022 10:59 AM To: Derek Rayside (SE Director) [email protected] Cc: Patrick Lam [email protected]; Charles Clarke [email protected] Subject: Re: should SE drop accreditation? --- asks the dean of engineering I confess that I have drunk the Kool-aid and believe in accreditation. Not for the software that is built in SV, but for software in engineered products and safety-critical systems (Software is even harder to get correct than the work done in most other engineering disciplines).
I agree with Charlie that the problem is with the one-size-fits-all accreditation requirements, not with the need for accreditation.
I also wonder what the difference between the SE and CS programs (or CS with the SE Option) would be without accreditation.
My two cents, Jo
On Oct 16, 2022, at 9:34 AM, Charles Clarke <[email protected] mailto:[email protected]> wrote:
If we don't need accreditation, why do we need an SE program? The potential need for accreditation was (at least partly) the reason for creating an SE program in the first place.
Personally, I think we do need accreditation. We need "proper engineers" who understand software, including AI. Maybe it's not so fashionable right now, but that will change, possibly with the growing impact of AI. I think the real problem is that accreditation is a crazy process that doesn't reflect reality. The Canadian Engineering Deans need to be working on that problem, which impacts all programs.
*From:*Derek Rayside (SE Director) <[email protected] mailto:[email protected]> *Sent:*Sunday, October 16, 2022 8:03 AM *To:*Joanne Atlee <[email protected] mailto:[email protected]>; Charles Clarke <[email protected] mailto:[email protected]>; Patrick Lam <[email protected] mailto:[email protected]> *Subject:*should SE drop accreditation? --- asks the dean of engineering Hello Jo, Charlie, and Pat,
*Should SE drop accreditation?*I am passing along this question from a surprising source: the dean of engineering. Context: I have been interviewing for the ADUG position, and so have had several long conversations with the dean. She asked this off the cuff, and might not have considered it deeply. She is very motivated by the concept of "curriculum diet" and improving student experience, so that is probably the lens she was applying.
(FYI: I will not be the next ADUG. With the combination of trying to support Stephanie's science, young kids, commute, plus the current engineering undergraduate office staffing situation, it is not logistically feasible for me to do that job. The selected candidate will do a good job and I'm not unhappy about the process.)
Historically I have been in favour of accreditation, if only for sentimental reasons.
Thinking practically instead of sentimentally/philosophically, what would we change if we didn't have to meet accreditation?
*1. Reduce natural sciences.*SE students do not need 5 courses in natural sciences. (FYI: I am working on overhauling this part of the curriculum within the bounds of accreditation. So still 5 courses, but more flexible in content and scheduling.)
*2. More choice?*The choices that students would want would likely be to dodge hard courses like CS343 and SE380. I'm not sure we would want to remove those. Perhaps we could have more of "pick 3 of these 4" kind of choices. (FYI: I am working on creating more scheduling flexibility without drastic changes to the set of courses required --- so within the bounds of accreditation.)
*3. Add more digital hardware?*Perhaps a computer architecture course, or a VLSI course? FYI: ECE is really short on instructors in this area now and until they can hire fast enough to replace retirements and long term medical leaves. So ECE would be opposed to us adding these courses in the next few years.
I guess my current view is that I'm already working on tweaking the curriculum to address these directions within the bounds of accreditation, so I'd like to see the dust settle on that first, which will be around the time of our next accreditation cycle ~2025/6. So I'm not personally trying to move in this direction at this time. But if the three of you were unanimously strongly in favour of moving away from accreditation, this might be a rare moment in history when the dean of engineering might agree to that, and I would not prioritize my sentimentality above the tide of history.
thanks, Derek.
More:
I don't know how risk adverse Mary is, but probably the right approach to making changes at the CEAB level is to do what we think is best and then to argue about it once they say no. If they say no to cognitive psychology, then you have a specific thing to argue about and at that point you get the other Deans to back you up. It probably would be good for the Deans to all agree in advance that they will back each other up on the things they want to change, which is why this plan needs Mary on side.
Personally, I think theoretical computer science is math, especially these days. IMHO, the correct route is to advocate for more math and less NSE in the SE program, rather than trying to argue that math is actually NSE.
--Charlie
From: Derek Rayside (SE Director) [email protected] Sent: Monday, October 17, 2022 9:03 AM To: Joanne Atlee [email protected] Cc: Patrick Lam [email protected]; Charles Clarke [email protected] Subject: Re: should SE drop accreditation? --- asks the dean of engineering
Cognitive science: I am thrilled to report that the 2023 calendar that will be published in January/February will include cognitive psychology courses as natural science electives. That was surprisingly easy to do. I really thought it was going to be a huge fight, but it went through without almost any discussion at all. Management Engineering had already opened this door a crack. I got ECE to walk through it with us. So at least three programs have cognitive psychology courses as approved natural science electives. NSERC funds cognitive psychology research, so I feel confident that this will be ok at the next accreditation.
Bioinformatics: At Engineering's FOPS committee last Friday (which discusses but doesn't vote), there was general acceptance of the following as natural science electives --- with some discussion, in which I argued that one of the major ways in which biology happens these days is by applying engineering/computing ideas towards natural science. Also, at least BIOL382 here involves reading actual scientific research papers and trying to reproduce/extend the techniques.
AMATH/BIOL382 Computational Modelling of Cellular Systems
CS482 Computational Techniques in Biological Sequence Analysis
PHYS468 Introduction to the Implementation of Quantum Information Processing
We have already let a few students take AMATH382 and PHYS468 on an experimental basis to learn more about the courses, and they really enjoyed them. It's pretty exciting for (at least some of) our students to be able to use their software skills to extend some published biology paper.
Hopefully we will be getting these bioinformatics courses (and more in this kind of area) in for the 2024 calendar. They might generate discussion at the next accreditation, but I'm hopeful that any qualified SE accreditation visitor will understand this.
Theoretical computer science: I would love to have this included in NSEs. Information theory too. This is a good thing to advocate to Engineers Canada for. I would love CS360 as an NSE. That would make the program better.
thanks! Derek.
On 10/17/22 08:06, Joanne Atlee wrote:
Whew! I am relieved by your summary note.
With respect to accreditation improvements and NSEs, one direction would be to push to reduce NSE requirements in favour of recognizing engineering science. There are aspects of theoretical computer science (complexity theory, limits of computability) that would be nice to count as science underlying SE.
Another direction would be to try to increase the boundaries of natural science — to include, for example, Cognitive Psychology, given that for a large percentage of software applications, the “environment” of the system are human users and in many cases a system’s success requires the cooperation and interaction of human operators. In the early days, SE Option students and SE degree students often tried to earn a Cog Psych Option as well.
Best, Jo
On Oct 17, 2022, at 7:31 AM, Derek Rayside (SE Director) [email protected] wrote:
Thank you all.
So I think we're on the same page here:
- SE should remain accredited.
- Losing accreditation would probably threaten the long-term viability of the program.
- Some students are interested in the iron ring.
- The accreditation process and requirements are a pain, and Engineering Deans across Canada should work on trying to fix it.
I think it's good for us to have had this little discussion to affirm our position on this question. In my observation, over the last 5-6 years, this is a question that comes up ad hoc at SE Board meetings etc, and then it kind of peters out, but it's always a question that is asked and answered by people who are outside of the actual operations of the program. I'm glad to have clear direction from inside the program on this question. Thank you.
Accreditation improvements? Is there anything we might like to suggest to the faculty of engineering about accreditation process and criteria improvements we'd like to see? Besides the obvious unnecessary friction the process has, which affects all engineering programs. Sometimes I think that it would be nice if we could reduce NSE requirements in favour of something else. The distinction between "natural science" and "engineering science" is also a point of friction for some other engineering programs (e.g., chemical has said this to me recently).
Digital Hardware: I was quite surprised at the ECE department retreat this summer when I asked some digital hardware folks if they thought SE students should have a choice between control theory and computer architecture, and they said no, control theory all the way. (Besides which, ECE could not staff teaching the SE students more digital hardware as a required course.) As Pat mentioned, I think more digital hardware might be good in the SE degree, but the ECE staffing situation has changed for the worse in this area over the last few years --- so, as Jim Barby sometimes says, we can ask this question again in 5-10 years.
I'm not sure if I misunderstood part of what Pat wrote below, but there is no change to the digital hardware content in the BSE degree. In particular, it's not reducing. What I am trying work with the SE CC on moving away from is so much overwhelming natural science in first year. This is a problem that Pat was also exploring during his directorship. Some things have changed in the external world since then (highschool math is weaker), and some other things have become clearer (many incoming students do not have the necessary math background for university physics). So the SE CC will look at swapping chemistry to 1A (like almost every other engineering program) and physics to 2A (after they have some university math skills).
Control theory: I am working with an ad hoc committee of alumni who have done further study in controls, including Rollen D'Souza, to redesign the labs for SE380. Additionally, Patrick recently connected with Bilal Akhtar from SE2019 who told us that the latest Go programming language garbage collector is now based on a PI controller. We're trying to build up a set of these kinds of examples so that we can make these courses more obviously relevant to SE students.
Thank you, Derek.
On 10/16/22 22:20, Patrick Lam wrote:
Accreditation is a big pain in the ass, as we all know. I agree with many of the other thoughts. Here's my calculations. I have no sentimental attachment to being accredited and if we could offer a good program without accreditation then we should.
- Accreditation forces us to have stuff in our curriculum that we don't want. There is no natural champion for, in particular, the NSEs.
- Students have a strong identification with "engineering undergraduate culture" (e.g. EngSoc) even if many students do end up switching to CS especially after first year. Being iron ring eligible is important to some subset of our students.
- SE has a distinct identity from CS, in a much stronger way than the SE option. To expand on what I said above, I think the principal drivers of that are the cohort system and the iron ring.
- I don't think actual curriculum content is that important to students, and I think it doesn't strongly affect what graduates are able to do 5 years down the road. My assumption is that a graduate can learn what they need to in terms of technical skills. Heresy alert: I'm skeptical that SE 463/464/465 contribute significantly to our students designing and writing better systems later on. On the other hand, the mandatory 3-course FYDP sequence does, I think, cause many students to stretch themselves in ways that they wouldn't otherwise.
We do talk about ethics perhaps a bit more than CS students get, but I also wouldn't put a lot of money on graduates internalizing engineers' responsibility to society. Well, I'd put $1000 on it, but I wouldn't put half a year's salary on it.
OK, so what if we had a cohorted program which had no iron ring? What would I tell people at OUF? Well, I could say the words about SE 463/464/465, but I wouldn't. I could say words about ECE 124 and ECE 222 and getting a bit more Computer Engineering education. We're moving away from them getting as much technical depth as 1st year computer engineers in hardware topics, so that's fine, but the words wouldn't be quite as accurate. I think the words I would believe most would be about FYDP and about the cohort system. It would definitely be more of a challenge to be convincing. We probably would still get enough students.
I think, by the way, that students would duck SE380, but I don't think the most academically inclined top 75% would duck CS 343. Sometimes I'm wrong about these things.
Derek, you've said previously that CS + DHW is a better SE degree than CS. We could move that way if we weren't accredited.
The last question is whether we'd need to remove "engineering" from the name.
To sum up, though, I kind of have a "why change it if it's not broken" attitude towards this question. We could probably make an unaccredited program work, but it's not at all clear that it would be long-term viable, and it's also not clear that it would help students, except that we can remove some NSEs. And we'd lose some things that are important to some students.
[I still want to write the SE retrospective, but writing papers and grants that are in my core research area keep on getting in the way; hopefully next month.]
pat
On 2022-10-17 04:27, Charles Clarke wrote:
There's also the very practical consideration that without accreditation (i.e., the need for SE to be partially supported by an Engineering program) SE will likely melt away into CS. Mary seems to like the SE program, which continues to attract good students. If she wants it to continue to be an Engineering program, it needs to actually be an engineering program.
From: Joanne Atlee [email protected] Sent: Sunday, October 16, 2022 10:59 AM To: Derek Rayside (SE Director) [email protected] Cc: Patrick Lam [email protected]; Charles Clarke [email protected] Subject: Re: should SE drop accreditation? --- asks the dean of engineering I confess that I have drunk the Kool-aid and believe in accreditation. Not for the software that is built in SV, but for software in engineered products and safety-critical systems (Software is even harder to get correct than the work done in most other engineering disciplines).
I agree with Charlie that the problem is with the one-size-fits-all accreditation requirements, not with the need for accreditation.
I also wonder what the difference between the SE and CS programs (or CS with the SE Option) would be without accreditation.
My two cents, Jo
On Oct 16, 2022, at 9:34 AM, Charles Clarke <[email protected] mailto:[email protected]> wrote:
If we don't need accreditation, why do we need an SE program? The potential need for accreditation was (at least partly) the reason for creating an SE program in the first place.
Personally, I think we do need accreditation. We need "proper engineers" who understand software, including AI. Maybe it's not so fashionable right now, but that will change, possibly with the growing impact of AI. I think the real problem is that accreditation is a crazy process that doesn't reflect reality. The Canadian Engineering Deans need to be working on that problem, which impacts all programs.
*From:*Derek Rayside (SE Director) <[email protected] mailto:[email protected]> *Sent:*Sunday, October 16, 2022 8:03 AM *To:*Joanne Atlee <[email protected] mailto:[email protected]>; Charles Clarke <[email protected] mailto:[email protected]>; Patrick Lam <[email protected] mailto:[email protected]> *Subject:*should SE drop accreditation? --- asks the dean of engineering Hello Jo, Charlie, and Pat,
*Should SE drop accreditation?*I am passing along this question from a surprising source: the dean of engineering. Context: I have been interviewing for the ADUG position, and so have had several long conversations with the dean. She asked this off the cuff, and might not have considered it deeply. She is very motivated by the concept of "curriculum diet" and improving student experience, so that is probably the lens she was applying.
(FYI: I will not be the next ADUG. With the combination of trying to support Stephanie's science, young kids, commute, plus the current engineering undergraduate office staffing situation, it is not logistically feasible for me to do that job. The selected candidate will do a good job and I'm not unhappy about the process.)
Historically I have been in favour of accreditation, if only for sentimental reasons.
Thinking practically instead of sentimentally/philosophically, what would we change if we didn't have to meet accreditation?
*1. Reduce natural sciences.*SE students do not need 5 courses in natural sciences. (FYI: I am working on overhauling this part of the curriculum within the bounds of accreditation. So still 5 courses, but more flexible in content and scheduling.)
*2. More choice?*The choices that students would want would likely be to dodge hard courses like CS343 and SE380. I'm not sure we would want to remove those. Perhaps we could have more of "pick 3 of these 4" kind of choices. (FYI: I am working on creating more scheduling flexibility without drastic changes to the set of courses required --- so within the bounds of accreditation.)
*3. Add more digital hardware?*Perhaps a computer architecture course, or a VLSI course? FYI: ECE is really short on instructors in this area now and until they can hire fast enough to replace retirements and long term medical leaves. So ECE would be opposed to us adding these courses in the next few years.
I guess my current view is that I'm already working on tweaking the curriculum to address these directions within the bounds of accreditation, so I'd like to see the dust settle on that first, which will be around the time of our next accreditation cycle ~2025/6. So I'm not personally trying to move in this direction at this time. But if the three of you were unanimously strongly in favour of moving away from accreditation, this might be a rare moment in history when the dean of engineering might agree to that, and I would not prioritize my sentimentality above the tide of history.
thanks, Derek.
--
Derek Rayside, PhD, PEng (he/him) Director, Software Engineering University of Waterloo Calendar
I respectfully acknowledge that the University of Waterloo is situated on the Haldimand Tract, land that was promised to the Haudenosaunee of the Six Nations of the Grand River, and is within the territory of the Neutral, Anishinaabe, and Haudenosaunee peoples.
--
Derek Rayside, PhD, PEng (he/him) Director, Software Engineering University of Waterloo Calendar
I respectfully acknowledge that the University of Waterloo is situated on the Haldimand Tract, land that was promised to the Haudenosaunee of the Six Nations of the Grand River, and is within the territory of the Neutral, Anishinaabe, and Haudenosaunee peoples.