- Physics-Departmental-Computing

Fwd: CHG0034304 - Office 365 License Update
by David Colling 03 Aug '16

03 Aug '16

FYI... -------- Forwarded Message -------- Subject: CHG0034304 - Office 365 License Update Date: Wed, 3 Aug 2016 16:47:36 +0000 From: ICT Service Desk <service.desk(a)imperial.ac.uk> To: ICTAP Members <ccs-dept(a)imperial.ac.uk> Office 365 License Update: Over the next few days ICT are updating Imperial College’s Office 365 licensing to simplify its management. Most staff and students will see no change. However, Honorary and Casual staff will gain the right to download Office 365 for home machines and mobile devices. Some users may lose access to Office 365 for a few seconds while the change is implemented. Should this persist they should contact the ICT Service Desk who will be able to advise accordingly. ICT will let you know once this change has completed, which we expect to be early next week.

1 0

FW: Print Service Survey
by Vine, Harry A 15 Jul '16

15 Jul '16

Dear Colleagues, If you wish please complete the ICT survey for the central touch-card print service or feed back any observations to me which I shall forward on. Particularly consider our teaching lab printer provision. For your information on the cost element; collectively our teaching lab printers generate a modest profit for Physics. Kind regards Harry Harry Vine Email: h.vine(a)imperial.ac.uk<mailto:h.vine@imperial.ac.uk> Departmental Services Manager 216a Blackett Laboratory Department of Physics Tel: +44 (0)20 7594 7515 South Kensington Campus Fax: +44 (0)20 7594 1441 Imperial College London SW7 2AZ, UK Web : http://www.imperial.ac.uk/ From: Allanson, Chris Sent: 12 July 2016 17:35 To: Subject: Print Service Survey Dear All The College is working on improving the central touch-card print service and needs your input to identify the areas that are currently working well and those needing improvement. So far ICT has identified 3 main areas to focus on: 1 Reliability of central services and all printers 2 Account, credit and allowance administration 3 Supply chain for consumables Please could you tell us via this short questionnaire about your experiences with the print service, what works well and what other improvements are required: https://imperial.eu.qualtrics.com/SE/?SID=SV_cMv00edyeskgdet Alternatively you can tell us your views by emailing myself, Alex Yakimov<mailto:a.yakimov@imperial.ac.uk> or your Digital Partner. Thank you for your contribution to improving the print service at Imperial College. Kind regards Chris Business Analyst for ICT

1 0

Fwd: Major Incident - INC0087301 - Phishing email with subject "Imperial Upgrade Outlook 2016"
by David Colling 13 Jul '16

13 Jul '16

-------- Forwarded Message -------- Subject: Major Incident - INC0087301 - Phishing email with subject "Imperial Upgrade Outlook 2016" Date: Wed, 13 Jul 2016 06:25:56 -0700 From: ICT Incident Manager <incident.manager(a)imperial.ac.uk> Reply-To: service.desk(a)imperial.ac.uk To: ict-major-incident-dl(a)imperial.ac.uk Dear Customer, The notes below refer to Major Incident INC0087301. A number of users have received a phishing email with the subject line "Imperial Upgrade Outlook 2016". The link in the email has now been blocked on the college network. If you have clicked on the link and entered your college credentials please call the ICT Service Desk on 02075949000. You can follow progress and future updates via the Service Status page here: https://imperial.service-now.com/ict/status.do For any comments relating to either the incident or this email, please contact ICT Incident Manager. Thank you, ICT Imperial College London Tel: +44 (0)20 7594 9000 (49000 internal) Web: imperial.ac.uk/admin-services/ict/contact-ict-service-desk Twitter: @TheServiceDesk ------------------------------------------------------------------------ You are currently subscribed to receive ICT Major Incident notifications. To unsubscribe, please email ICT Incident Manager. Ref:MSG5009587

1 0

Re: [Physics-Departmental-Computing] Single Python build - survey results
by David Colling 12 Jul '16

12 Jul '16

Thanks Martin, BTW will the student machines be Windows 10 by October? Best, david On 12/07/16 10:28, Grune, Martin wrote: > I've not got anything ready for a test yet, but I'm working on it. I'll let you know as soon as I have something for you to try. > > -----Original Message----- > From: David Colling [mailto:d.colling@imperial.ac.uk] > Sent: 11 July 2016 22:03 > To: Grune, Martin <m.grune(a)imperial.ac.uk>; Bennett, Will <w.bennett(a)imperial.ac.uk> > Cc: Khanna, Taru <t.khanna(a)imperial.ac.uk>; Bennett, Jason W <jason.bennett(a)imperial.ac.uk>; physics-departmental-computing <physics-departmental-computing(a)imperial.ac.uk> > Subject: Re: Single Python build - survey results > > Thanks Martin, > > This sounds very good. Please can I have a windows VM with the windows teaching environment and Anaconda installed for tests? I suspect that this will be a good solution looking at the Anaconda website. > > Best, > david > > On 11/07/16 17:59, Grune, Martin wrote: >> Hi David >> >> I've just tested how 2.7 and 3.5 mix and it's pretty simple in Anaconda. I assume you use ipython for a lot of your work? After the environment is created, you simply type "activate python2" and you're in python 2.7, "deactivate" brings you back to 3.5. Modules seem to be fairly well supported too, there are one or two issues but I think we can resolve them. Either way, I get the feeling anaconda is going to be a good selection. >> >> Regards >> >> Martin >> >> -----Original Message----- >> From: David Colling [mailto:d.colling@imperial.ac.uk] >> Sent: 11 July 2016 17:52 >> To: Bennett, Will <w.bennett(a)imperial.ac.uk>; Grune, Martin >> <m.grune(a)imperial.ac.uk> >> Cc: Khanna, Taru <t.khanna(a)imperial.ac.uk>; Bennett, Jason W >> <jason.bennett(a)imperial.ac.uk>; physics-departmental-computing >> <physics-departmental-computing(a)imperial.ac.uk> >> Subject: Re: Single Python build - survey results >> >> Hi Will, >> >> Well it looks like we should go with anaconda, have both 2.7 and 3.5 installed but have 3.5 as the default. My vote for 2.7 was only because I have to deal with legacy stuff. >> >> I have cc'ed the physics computing list in case anybody else has any comments. >> >> Also I found a pge about the differences between the two and this can be found at: >> >> http://sebastianraschka.com/Articles/2014_python_2_3_key_diff.html >> >> Which be useful for those preparing teaching materials. >> >> Please don't update the system in Blackett until after the Global Summer School on the week of the 8th. >> >> Best, >> david >> >> On 11/07/16 16:37, Bennett, Will wrote: >>> Good afternoon all, >>> >>> >>> >>> Please find the key points and comments from the survey below. In >>> summary, it seems that most people would like both versions if >>> possible, and Anaconda got the most positive response for >>> distribution system and everyone wants Jupyter. I've also >>> consolidated the list of requested modules below as well - hopefully >>> these can be included in the packaging (if not already in Anaconda). >>> >>> >>> >>> David - can you let me know your thoughts? >>> >>> >>> >>> Many thanks >>> >>> Will. >>> >>> >>> >>> >>> >>> >>> >>> *__* >>> >>> *_Which version of Python_* >>> >>> >>> >>> * Version 2.7 - 8 votes (47%) >>> >>> * Version 3 - 9 votes (53%) >>> >>> >>> >>> Key Comments: >>> >>> >>> >>> Most of the world still uses Python 2.7 even though Python 3 is a >>> better language. I think that both should be available. >>> >>> Version 3 is the one going forward, and due to syntax >>> incompatibilities is the one we based our course on exclusively. >>> >>> Python 3+ is the future of the language. Almost all scientific >>> packages are now supported. 3.5 in particular adds a dot product >>> operator which simplifies mathematical syntax. >>> >>> Both versions should be available. >>> >>> I am making the switch this year from 2.7 to 3 as I feel a sufficient >>> number of the science and engineering packages I use have made the switch. >>> >>> Above option does not allow us to state both versions. I think it >>> would be good to have both versions. On our local systems (Linux and >>> Mac OS >>> X) I always install both versions. As 3 is now required for some work. >>> >>> Even though 2.7 is the more widely used Python, it has been out of >>> active development for a long time and we should be forward looking >>> in what we teach. >>> >>> Python 2.7 has a number of design problems and inconsistencies that >>> are not going to be fixed. >>> >>> Many of these problems are resolved in Python 3, which is >>> consequently more logical and consistent. >>> >>> Although Python 2.7 (and even 2.6) is still widely used for backward >>> compatibility with legacy code. However, Python 3 use is gradually >>> replacing Python 2. Python 3 has now been around for about a decade >>> and is stable and supported - I can't at the moment think of any >>> libraries that are not available to Python 3 now. >>> >>> >>> >>> For most of our students this is their first introduction to >>> programming. Logical consistency in the language is very important. >>> In that context, Python 3 has significant advantages. >>> >>> Also, since this is pretty much their first time using Python (and >>> indeed programming at all) the issues of interfacing with >>> >>> their legacy code are likely to be at their least. In short, it is a >>> good time for the students to start with Python 3. >>> >>> >>> >>> That being said, there is still a need for Python 2.7 for those who >>> are already using it and who do still need to access legacy code. >>> >>> >>> >>> Therefore, my clear preference would be for Python3 to be the >>> default, but for a full offering of Python 2.7 and libraries to be available too. >>> >>> Python 2.7 is still the version of choice at many international user >>> facilities, and comes preinstalled on Mac and many flavours of Linux. >>> >>> >>> >>> I would support having both 2.7 and 3.5 installed on all machines. >>> >>> No strong feelings. But we do have most of our code 2.7 compatible >>> currently. >>> >>> >>> >>> >>> >>> >>> >>> *_Which Deployment Manager?_* >>> >>> >>> >>> * Anaconda - 10 votes (67%) >>> >>> * Canopy - 5 votes (33%) >>> >>> >>> >>> Key Comments: >>> >>> >>> >>> I prefer the interface. (Canopy) >>> >>> No strong feeling (Canopy) >>> >>> No particular reason, it has worked well for us. (Anaconda) >>> >>> Anaconda has taken over as the most popular manager. It's much easier >>> to package binaries up using conda than canopy, and it trivially >>> allows multiple different python distributions (if for instance some >>> need 2.7 and other 3.5) >>> >>> No preference >>> >>> I have been getting my students to use both over the last few years >>> to get a feeling for the strengths and weaknesses both. I now >>> strongly favour anaconda because I have found it more stable for >>> ipython notebooks and fonts from mathematical formula are better >>> rendered via >>> ipython+chrome than with Canopy - where often the maths fonts were >>> nearly unreadable. >>> >>> Afraid only used on Linux and Mac OS X. We switched from Canopy to >>> Anaconda as it was much easier to install and maintain. If I also >>> recall correctly, at that time canopy did not offer version 3. WE >>> have found anaconda so easy to work with (eg install packages) that >>> we would not now switch. >>> >>> This is more of a negative vote against Canopy. Canopy is >>> sufficiently clunky that most of my class ended up using their own laptops. >>> >>> I don't have a firm view on this binary choice. (I don't actually use >>> either for my own work, preferring the (in my view, superior) package >>> management of a GNU/Linux distribution.) In Physics, we have used >>> Canopy for our UG teaching. It is ok, but has had some irritating >>> issues. I have not used Anaconda, but from my brief reading around it >>> would appear to offer both Python 2 and Python 3 and the libraries >>> that I would view as important. However, having not used it myself, >>> it would be a bit of a leap in the dark. >>> >>> I have regularly used Canopy when teaching and in my own research, >>> and I am very happy with it. >>> >>> I cannot justify this choice, as I've only used Canopy. I'm sure >>> either would be fine. >>> >>> That's what we are used to... also, anaconda supports both Python 2 >>> and >>> 3 environments >>> >>> >>> >>> >>> >>> >>> >>> *_Include Jupyter?_* >>> >>> >>> >>> * Yes - 13 votes (93%) >>> >>> * No - 0 votes (0%) >>> >>> * Don't Have an opinion - 1 vote (7%) >>> >>> >>> >>> Key Comments: >>> >>> >>> >>> None >>> >>> >>> >>> >>> >>> *_Key Modules to include:_* >>> >>> >>> >>> Modules Identified for inclusion in Python build: >>> >>> >>> >>> * Scipy >>> >>> * Numpy >>> >>> * Matplotlib >>> >>> * Networkx >>> >>> * iPython >>> >>> * RDKit >>> >>> * /Everything in default Anaconda distribution/ >>> >>> * Bioconductor >>> >>> * Virtualenv (Virtual Environments) >>> >>> * Pytest >>> >>> * Sympy >>> >>> * Cython >>> >>> * Pandas >>> >>> * Sphynx >>> >>> * Tkinter >>> >>> * subprocess >>> >>> >>> >>> >>> >>> *_Other Comments:_* >>> >>> >>> >>> It needs to be supported on Windows, Mac and Linux and needs to be >>> free for our students and staff. >>> >>> Previously I have requested both Enthought and Anaconda to be >>> installed on ESE clusters for teaching. This year I will only be requesting anaconda. >>> >>> Please note that it would be good to have both V2 and V3 of python >>> >>> In order for this to be useful for my module (and, I would hope, >>> others) we would also need Git installed in all the labs. Revision >>> control is an indispensable part of software engineering and there is >>> little point in providing students with a programming environment if >>> they don't also have access to revision control software. >>> >>> Since we are already making extensive use of a Python distro (Canopy) >>> in our teaching, it is essential that we have sufficient time to >>> review thoroughly whatever is proposed in good time for the start of >>> the next academic year It would be a serious problem if we ended up >>> with a system that removed or broke features that we rely on in our >>> existing teaching without time to fix (or even in the worst-case, >>> roll >>> back!) >>> >>> >>> >>> >>> >>> *_Key Contacts to include:_* >>> >>> >>> >>> d.colling(a)imperial.ac.uk >>> >>> malhado(a)imperial.ac.uk >>> >>> Gerard Gorman >>> >>> suhail.islam(a)imperial.ac.uk >>> >>> david.ham(a)imperial.ac.uk >>> >>> carl(a)imperial.ac.uk >>> >>> s.pawar(a)imperial.ac.uk >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> *Will Bennett* >>> >>> Digital Partner (Faculty of Natural Sciences) >>> >>> Information & Communication Technologies >>> >>> Level 4 Sherfield >>> >>> Imperial College London >>> >>> South Kensington Campus >>> >>> London SW7 2AZ >>> >>> >>> >>> Mob: 07872 850 035 >>> >>> Tel: 0207 59 49277 >>> >>> Email: w.bennett(a)imperial.ac.uk <mailto:w.bennett@imperial.ac.uk> >>> >>> http://www.imperial.ac.uk/admin-services/ict/ >>> >>> >>>

1 0

Re: [Physics-Departmental-Computing] Single Python build - survey results
by David Colling 11 Jul '16

11 Jul '16

Thanks Martin, This sounds very good. Please can I have a windows VM with the windows teaching environment and Anaconda installed for tests? I suspect that this will be a good solution looking at the Anaconda website. Best, david On 11/07/16 17:59, Grune, Martin wrote: > Hi David > > I've just tested how 2.7 and 3.5 mix and it's pretty simple in Anaconda. I assume you use ipython for a lot of your work? After the environment is created, you simply type "activate python2" and you're in python 2.7, "deactivate" brings you back to 3.5. Modules seem to be fairly well supported too, there are one or two issues but I think we can resolve them. Either way, I get the feeling anaconda is going to be a good selection. > > Regards > > Martin > > -----Original Message----- > From: David Colling [mailto:d.colling@imperial.ac.uk] > Sent: 11 July 2016 17:52 > To: Bennett, Will <w.bennett(a)imperial.ac.uk>; Grune, Martin <m.grune(a)imperial.ac.uk> > Cc: Khanna, Taru <t.khanna(a)imperial.ac.uk>; Bennett, Jason W <jason.bennett(a)imperial.ac.uk>; physics-departmental-computing <physics-departmental-computing(a)imperial.ac.uk> > Subject: Re: Single Python build - survey results > > Hi Will, > > Well it looks like we should go with anaconda, have both 2.7 and 3.5 installed but have 3.5 as the default. My vote for 2.7 was only because I have to deal with legacy stuff. > > I have cc'ed the physics computing list in case anybody else has any comments. > > Also I found a pge about the differences between the two and this can be found at: > > http://sebastianraschka.com/Articles/2014_python_2_3_key_diff.html > > Which be useful for those preparing teaching materials. > > Please don't update the system in Blackett until after the Global Summer School on the week of the 8th. > > Best, > david > > On 11/07/16 16:37, Bennett, Will wrote: >> Good afternoon all, >> >> >> >> Please find the key points and comments from the survey below. In >> summary, it seems that most people would like both versions if >> possible, and Anaconda got the most positive response for distribution >> system and everyone wants Jupyter. I've also consolidated the list of >> requested modules below as well - hopefully these can be included in >> the packaging (if not already in Anaconda). >> >> >> >> David - can you let me know your thoughts? >> >> >> >> Many thanks >> >> Will. >> >> >> >> >> >> >> >> *__* >> >> *_Which version of Python_* >> >> >> >> * Version 2.7 - 8 votes (47%) >> >> * Version 3 - 9 votes (53%) >> >> >> >> Key Comments: >> >> >> >> Most of the world still uses Python 2.7 even though Python 3 is a >> better language. I think that both should be available. >> >> Version 3 is the one going forward, and due to syntax >> incompatibilities is the one we based our course on exclusively. >> >> Python 3+ is the future of the language. Almost all scientific >> packages are now supported. 3.5 in particular adds a dot product >> operator which simplifies mathematical syntax. >> >> Both versions should be available. >> >> I am making the switch this year from 2.7 to 3 as I feel a sufficient >> number of the science and engineering packages I use have made the switch. >> >> Above option does not allow us to state both versions. I think it >> would be good to have both versions. On our local systems (Linux and >> Mac OS >> X) I always install both versions. As 3 is now required for some work. >> >> Even though 2.7 is the more widely used Python, it has been out of >> active development for a long time and we should be forward looking in >> what we teach. >> >> Python 2.7 has a number of design problems and inconsistencies that >> are not going to be fixed. >> >> Many of these problems are resolved in Python 3, which is consequently >> more logical and consistent. >> >> Although Python 2.7 (and even 2.6) is still widely used for backward >> compatibility with legacy code. However, Python 3 use is gradually >> replacing Python 2. Python 3 has now been around for about a decade >> and is stable and supported - I can't at the moment think of any >> libraries that are not available to Python 3 now. >> >> >> >> For most of our students this is their first introduction to >> programming. Logical consistency in the language is very important. >> In that context, Python 3 has significant advantages. >> >> Also, since this is pretty much their first time using Python (and >> indeed programming at all) the issues of interfacing with >> >> their legacy code are likely to be at their least. In short, it is a >> good time for the students to start with Python 3. >> >> >> >> That being said, there is still a need for Python 2.7 for those who >> are already using it and who do still need to access legacy code. >> >> >> >> Therefore, my clear preference would be for Python3 to be the >> default, but for a full offering of Python 2.7 and libraries to be available too. >> >> Python 2.7 is still the version of choice at many international user >> facilities, and comes preinstalled on Mac and many flavours of Linux. >> >> >> >> I would support having both 2.7 and 3.5 installed on all machines. >> >> No strong feelings. But we do have most of our code 2.7 compatible >> currently. >> >> >> >> >> >> >> >> *_Which Deployment Manager?_* >> >> >> >> * Anaconda - 10 votes (67%) >> >> * Canopy - 5 votes (33%) >> >> >> >> Key Comments: >> >> >> >> I prefer the interface. (Canopy) >> >> No strong feeling (Canopy) >> >> No particular reason, it has worked well for us. (Anaconda) >> >> Anaconda has taken over as the most popular manager. It's much easier >> to package binaries up using conda than canopy, and it trivially >> allows multiple different python distributions (if for instance some >> need 2.7 and other 3.5) >> >> No preference >> >> I have been getting my students to use both over the last few years to >> get a feeling for the strengths and weaknesses both. I now strongly >> favour anaconda because I have found it more stable for ipython >> notebooks and fonts from mathematical formula are better rendered via >> ipython+chrome than with Canopy - where often the maths fonts were >> nearly unreadable. >> >> Afraid only used on Linux and Mac OS X. We switched from Canopy to >> Anaconda as it was much easier to install and maintain. If I also >> recall correctly, at that time canopy did not offer version 3. WE >> have found anaconda so easy to work with (eg install packages) that we >> would not now switch. >> >> This is more of a negative vote against Canopy. Canopy is sufficiently >> clunky that most of my class ended up using their own laptops. >> >> I don't have a firm view on this binary choice. (I don't actually use >> either for my own work, preferring the (in my view, superior) package >> management of a GNU/Linux distribution.) In Physics, we have used >> Canopy for our UG teaching. It is ok, but has had some irritating >> issues. I have not used Anaconda, but from my brief reading around it >> would appear to offer both Python 2 and Python 3 and the libraries >> that I would view as important. However, having not used it myself, it >> would be a bit of a leap in the dark. >> >> I have regularly used Canopy when teaching and in my own research, and >> I am very happy with it. >> >> I cannot justify this choice, as I've only used Canopy. I'm sure >> either would be fine. >> >> That's what we are used to... also, anaconda supports both Python 2 >> and >> 3 environments >> >> >> >> >> >> >> >> *_Include Jupyter?_* >> >> >> >> * Yes - 13 votes (93%) >> >> * No - 0 votes (0%) >> >> * Don't Have an opinion - 1 vote (7%) >> >> >> >> Key Comments: >> >> >> >> None >> >> >> >> >> >> *_Key Modules to include:_* >> >> >> >> Modules Identified for inclusion in Python build: >> >> >> >> * Scipy >> >> * Numpy >> >> * Matplotlib >> >> * Networkx >> >> * iPython >> >> * RDKit >> >> * /Everything in default Anaconda distribution/ >> >> * Bioconductor >> >> * Virtualenv (Virtual Environments) >> >> * Pytest >> >> * Sympy >> >> * Cython >> >> * Pandas >> >> * Sphynx >> >> * Tkinter >> >> * subprocess >> >> >> >> >> >> *_Other Comments:_* >> >> >> >> It needs to be supported on Windows, Mac and Linux and needs to be >> free for our students and staff. >> >> Previously I have requested both Enthought and Anaconda to be >> installed on ESE clusters for teaching. This year I will only be requesting anaconda. >> >> Please note that it would be good to have both V2 and V3 of python >> >> In order for this to be useful for my module (and, I would hope, >> others) we would also need Git installed in all the labs. Revision >> control is an indispensable part of software engineering and there is >> little point in providing students with a programming environment if >> they don't also have access to revision control software. >> >> Since we are already making extensive use of a Python distro (Canopy) >> in our teaching, it is essential that we have sufficient time to >> review thoroughly whatever is proposed in good time for the start of >> the next academic year It would be a serious problem if we ended up >> with a system that removed or broke features that we rely on in our >> existing teaching without time to fix (or even in the worst-case, roll >> back!) >> >> >> >> >> >> *_Key Contacts to include:_* >> >> >> >> d.colling(a)imperial.ac.uk >> >> malhado(a)imperial.ac.uk >> >> Gerard Gorman >> >> suhail.islam(a)imperial.ac.uk >> >> david.ham(a)imperial.ac.uk >> >> carl(a)imperial.ac.uk >> >> s.pawar(a)imperial.ac.uk >> >> >> >> >> >> >> >> >> >> >> >> *Will Bennett* >> >> Digital Partner (Faculty of Natural Sciences) >> >> Information & Communication Technologies >> >> Level 4 Sherfield >> >> Imperial College London >> >> South Kensington Campus >> >> London SW7 2AZ >> >> >> >> Mob: 07872 850 035 >> >> Tel: 0207 59 49277 >> >> Email: w.bennett(a)imperial.ac.uk <mailto:w.bennett@imperial.ac.uk> >> >> http://www.imperial.ac.uk/admin-services/ict/ >> >> >>

1 0

Re: [Physics-Departmental-Computing] Single Python build - survey results
by David Colling 11 Jul '16

11 Jul '16

Hi Will, Well it looks like we should go with anaconda, have both 2.7 and 3.5 installed but have 3.5 as the default. My vote for 2.7 was only because I have to deal with legacy stuff. I have cc'ed the physics computing list in case anybody else has any comments. Also I found a pge about the differences between the two and this can be found at: http://sebastianraschka.com/Articles/2014_python_2_3_key_diff.html Which be useful for those preparing teaching materials. Please don't update the system in Blackett until after the Global Summer School on the week of the 8th. Best, david On 11/07/16 16:37, Bennett, Will wrote: > Good afternoon all, > > > > Please find the key points and comments from the survey below. In > summary, it seems that most people would like both versions if possible, > and Anaconda got the most positive response for distribution system and > everyone wants Jupyter. I’ve also consolidated the list of requested > modules below as well – hopefully these can be included in the packaging > (if not already in Anaconda). > > > > David – can you let me know your thoughts? > > > > Many thanks > > Will. > > > > > > > > *__* > > *_Which version of Python_* > > > > · Version 2.7 - 8 votes (47%) > > · Version 3 - 9 votes (53%) > > > > Key Comments: > > > > Most of the world still uses Python 2.7 even though Python 3 is a better > language. I think that both should be available. > > Version 3 is the one going forward, and due to syntax incompatibilities > is the one we based our course on exclusively. > > Python 3+ is the future of the language. Almost all scientific packages > are now supported. 3.5 in particular adds a dot product operator which > simplifies mathematical syntax. > > Both versions should be available. > > I am making the switch this year from 2.7 to 3 as I feel a sufficient > number of the science and engineering packages I use have made the switch. > > Above option does not allow us to state both versions. I think it would > be good to have both versions. On our local systems (Linux and Mac OS > X) I always install both versions. As 3 is now required for some work. > > Even though 2.7 is the more widely used Python, it has been out of > active development for a long time and we should be forward looking in > what we teach. > > Python 2.7 has a number of design problems and inconsistencies that are > not going to be fixed. > > Many of these problems are resolved in Python 3, which is consequently > more logical and consistent. > > Although Python 2.7 (and even 2.6) is still widely used for backward > compatibility with legacy code. However, Python 3 use is gradually > replacing Python 2. Python 3 has now been around for about a decade and > is stable and supported - I can't at the moment think of any libraries > that are not available to Python 3 now. > > > > For most of our students this is their first introduction to > programming. Logical consistency in the language is very important. > In that context, Python 3 has significant advantages. > > Also, since this is pretty much their first time using Python (and > indeed programming at all) the issues of interfacing with > > their legacy code are likely to be at their least. In short, it is a > good time for the students to start with Python 3. > > > > That being said, there is still a need for Python 2.7 for those who are > already using it and who do still need to access legacy code. > > > > Therefore, my clear preference would be for Python3 to be the default, > but for a full offering of Python 2.7 and libraries to be available too. > > Python 2.7 is still the version of choice at many international user > facilities, and comes preinstalled on Mac and many flavours of Linux. > > > > I would support having both 2.7 and 3.5 installed on all machines. > > No strong feelings. But we do have most of our code 2.7 compatible > currently. > > > > > > > > *_Which Deployment Manager?_* > > > > · Anaconda - 10 votes (67%) > > · Canopy - 5 votes (33%) > > > > Key Comments: > > > > I prefer the interface. (Canopy) > > No strong feeling (Canopy) > > No particular reason, it has worked well for us. (Anaconda) > > Anaconda has taken over as the most popular manager. It's much easier to > package binaries up using conda than canopy, and it trivially allows > multiple different python distributions (if for instance some need 2.7 > and other 3.5) > > No preference > > I have been getting my students to use both over the last few years to > get a feeling for the strengths and weaknesses both. I now strongly > favour anaconda because I have found it more stable for ipython > notebooks and fonts from mathematical formula are better rendered via > ipython+chrome than with Canopy - where often the maths fonts were > nearly unreadable. > > Afraid only used on Linux and Mac OS X. We switched from Canopy to > Anaconda as it was much easier to install and maintain. If I also > recall correctly, at that time canopy did not offer version 3. WE have > found anaconda so easy to work with (eg install packages) that we would > not now switch. > > This is more of a negative vote against Canopy. Canopy is sufficiently > clunky that most of my class ended up using their own laptops. > > I don't have a firm view on this binary choice. (I don't actually use > either for my own work, preferring the (in my view, superior) package > management of a GNU/Linux distribution.) In Physics, we have used > Canopy for our UG teaching. It is ok, but has had some irritating > issues. I have not used Anaconda, but from my brief reading around it > would appear to offer both Python 2 and Python 3 and the libraries that > I would view as important. However, having not used it myself, it would > be a bit of a leap in the dark. > > I have regularly used Canopy when teaching and in my own research, and I > am very happy with it. > > I cannot justify this choice, as I've only used Canopy. I'm sure either > would be fine. > > That's what we are used to... also, anaconda supports both Python 2 and > 3 environments > > > > > > > > *_Include Jupyter?_* > > > > · Yes - 13 votes (93%) > > · No - 0 votes (0%) > > · Don't Have an opinion - 1 vote (7%) > > > > Key Comments: > > > > None > > > > > > *_Key Modules to include:_* > > > > Modules Identified for inclusion in Python build: > > > > · Scipy > > · Numpy > > · Matplotlib > > · Networkx > > · iPython > > · RDKit > > · /Everything in default Anaconda distribution/ > > · Bioconductor > > · Virtualenv (Virtual Environments) > > · Pytest > > · Sympy > > · Cython > > · Pandas > > · Sphynx > > · Tkinter > > · subprocess > > > > > > *_Other Comments:_* > > > > It needs to be supported on Windows, Mac and Linux and needs to be free > for our students and staff. > > Previously I have requested both Enthought and Anaconda to be installed > on ESE clusters for teaching. This year I will only be requesting anaconda. > > Please note that it would be good to have both V2 and V3 of python > > In order for this to be useful for my module (and, I would hope, others) > we would also need Git installed in all the labs. Revision control is an > indispensable part of software engineering and there is little point in > providing students with a programming environment if they don't also > have access to revision control software. > > Since we are already making extensive use of a Python distro (Canopy) in > our teaching, it is essential that we have sufficient time to review > thoroughly whatever is proposed in good time for the start of the next > academic year It would be a serious problem if we ended up with a > system that removed or broke features that we rely on in our existing > teaching without time to fix (or even in the worst-case, roll back!) > > > > > > *_Key Contacts to include:_* > > > > d.colling(a)imperial.ac.uk > > malhado(a)imperial.ac.uk > > Gerard Gorman > > suhail.islam(a)imperial.ac.uk > > david.ham(a)imperial.ac.uk > > carl(a)imperial.ac.uk > > s.pawar(a)imperial.ac.uk > > > > > > > > > > > > *Will Bennett* > > Digital Partner (Faculty of Natural Sciences) > > Information & Communication Technologies > > Level 4 Sherfield > > Imperial College London > > South Kensington Campus > > London SW7 2AZ > > > > Mob: 07872 850 035 > > Tel: 0207 59 49277 > > Email: w.bennett(a)imperial.ac.uk <mailto:w.bennett@imperial.ac.uk> > > http://www.imperial.ac.uk/admin-services/ict/ > > >

1 0

Fwd: FW: chemdraw licence / expiring
by David Colling 11 Jul '16

11 Jul '16

Any interests? -------- Forwarded Message -------- Subject: FW: chemdraw licence / expiring Date: Mon, 11 Jul 2016 08:27:03 +0000 From: Bennett, Will <w.bennett(a)imperial.ac.uk> To: Colling, David J <d.colling(a)imperial.ac.uk> CC: Giuseppe Mallia <malliagiuseppe(a)googlemail.com> Good morning David, Please find a message below from Giuseppe in Chemistry – they are keen to see if any other departments have an interest in ChemDraw (further details below). Could you please let either myself or Giuseppe (cc’d) know if Physics are interested in this? Many thanks, Will. *From:*Giuseppe Mallia [mailto:malliagiuseppe@googlemail.com] *Sent:* 11 July 2016 08:31 *To:* Bennett, Will <w.bennett(a)imperial.ac.uk> *Cc:* Parkins, Kat <k.parkins(a)imperial.ac.uk> *Subject:* chemdraw licence / expiring Dear Will (cc: Kat), we're in the process of renewing the licence for ChemDraw. There are two options: 1) ChemDraw Prime (GBP 2,057.00) 2) ChemDraw Professional (GBP3,878.00) The current license is going to expire at the end of July. Can you please forwards this email to the ICTrep in the Dept. of LF, Medicine, ChemEng, Materials and Physics to see if they have any interest? All the best, Giu ################################################### ChemDraw Prime is the complete entry-level drawing program that provides all the tools you need to quickly create chemically intelligent, publication-ready chemical structures and reactions, laboratory notes and experiment write-ups. As well as a full set of chemical structure essentials such as rings, bonds, chains, atoms and functional groups, ChemDraw Prime includes property calculators, chemical and lab equipment templates and handy TLC and Gel Electrophoresis Plate drawing tools. ChemDraw Professional is the complete drawing tool of choice for chemists and biologists to create publication-ready, scientifically intelligent drawings for use in ELNs, databases and publications and for querying chemical databases, now including SciFinder®. ChemDraw Professional includes all the features in ChemDraw Prime extended with biopolymer tools, BioDraw, NMR prediction and advanced Name to Structure. It also enables structure data exploration, seamless workflows with Elements, and supports organization and processing with ChemDraw for Excel, ChemFinder Standard, ChemScript and ChemDraw 3D. Please refer to the following webpage: www.cambridgesoft.com/software/overview.aspx?cid=4352&gclid=CjwKEAjw-_e7BRD… <http://www.cambridgesoft.com/software/overview.aspx?cid=4352&gclid=CjwKEAjw…> where the features of both versions are indicated. -- Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html

1 0

EPSRC Capital call
by David Colling 17 Jun '16

17 Jun '16

The following EPSRC call might be interesting to some of us: https://www.epsrc.ac.uk/funding/calls/tier2capitalinvestments/ I am pretty sure that ICT will haave picked this up but if they haven't yet then hopefully the ICT people on this list can forward it to them. Best, david

1 0

Fwd: International Integrated Population Microdata for the World - DSI Lunchtime Seminar
by David Colling 16 Jun '16

16 Jun '16

Sadly, I cannot make this lecture tomorrow but thought that some of you might find it interesting... -------- Forwarded Message -------- Subject: International Integrated Population Microdata for the World - DSI Lunchtime Seminar Date: Mon, 13 Jun 2016 13:14:40 +0000 From: McDonagh, Dominic <d.mcdonagh(a)imperial.ac.uk> To: Muggleton, Stephen H <s.muggleton(a)imperial.ac.uk>, Rodriguez Villegas, Esther <e.rodriguez(a)imperial.ac.uk>, Rueckert, Daniel <d.rueckert(a)imperial.ac.uk>, Matthews, Paul <p.matthews(a)imperial.ac.uk>, Yaliraki, Sophia <s.yaliraki(a)imperial.ac.uk>, Faisal, Aldo A <a.faisal(a)imperial.ac.uk>, Haigh, Joanna D <j.haigh(a)imperial.ac.uk>, Pain, Christopher <c.pain(a)imperial.ac.uk>, Jaffe, Andrew H <a.jaffe(a)imperial.ac.uk>, Leroi, Armand M <a.leroi(a)imperial.ac.uk>, Trotta, Roberto <r.trotta(a)imperial.ac.uk>, Stumpf, Michael P H <m.stumpf(a)imperial.ac.uk>, Heard, Nicholas A <n.heard(a)imperial.ac.uk>, Yeatman, Eric <e.yeatman(a)imperial.ac.uk>, Adams, Niall M <n.adams(a)imperial.ac.uk>, Knottenbelt, William J <w.knottenbelt(a)imperial.ac.uk>, Kosowski, Robert <r.kosowski(a)imperial.ac.uk>, Kennedy, Mark T <mark.kennedy(a)imperial.ac.uk>, Haskel, Jonathan <j.haskel(a)imperial.ac.uk>, Heinis, Thomas <t.heinis(a)imperial.ac.uk>, McCann, Julie A <j.mccann(a)imperial.ac.uk>, Shah, Nilay <n.shah(a)imperial.ac.uk>, Rosindell, James L <j.rosindell(a)imperial.ac.uk>, Polak, John W <j.polak(a)imperial.ac.uk>, Constantinides, George A <g.constantinides(a)imperial.ac.uk>, Hankin, Chris L <c.hankin(a)imperial.ac.uk>, Sternberg, Michael J E <m.sternberg(a)imperial.ac.uk>, Nicholson, Jeremy K <j.nicholson(a)imperial.ac.uk>, McCann, Julie A <j.mccann(a)imperial.ac.uk>, Toumazou, Christofer <c.toumazou(a)imperial.ac.uk>, Colling, David J <d.colling(a)imperial.ac.uk>, van Dyk, David A <d.van-dyk(a)imperial.ac.uk>, Rueckert, Daniel <d.rueckert(a)imperial.ac.uk>, Elliott, Paul <p.elliott(a)imperial.ac.uk>, Ezzati, Majid <majid.ezzati(a)imperial.ac.uk> Dear DSI Investigators, I’d like to invite you to our upcoming lunch seminar this Friday being delivered by Professor Robert McCaa (University of MinnesotaPopulation Center). DSI Lunchtime Seminar – Friday 17^th June, 13:00 – 14:00 *International Integrated Population Microdata for the World* Professor Robert McCaa (University of MinnesotaPopulation Center) Royal School of Mines LT 2.28 Further details here <http://www3.imperial.ac.uk/newsandeventspggrp/imperialcollege/engineering/d…> Complimentary lunch served. / / /The census microdata revolution, foretold over a decade ago, has arrived, but it is not yet complete. Then, census microdata were available for only a handful of countries and trans-border access was impossible for all but a few. Now, from the website www.ipums.org <http://www.ipums.org>, hosted by the University of Minnesota Population Center, many decades of census microdata for much of the world are readily accessible free of cost to bona fide researchers anywhere—regardless of country of birth, residence or citizenship./ / / /At this lunchtime seminar Professor McCaa from the University of Minnesota Population Center describes the IPUMS project, and the great potential of the data for researchers./ Best wishes, *Dominic McDonagh | *Business Development Officer Data Science Institute*|* Imperial College London d.mcdonagh(a)imperial.ac.uk <mailto:d.mcdonagh@imperial.ac.uk> +44 (0)20 7594 8291 www.imperial.ac.uk/data-science <http://www.imperial.ac.uk/data-science> @ImperialDSI <https://twitter.com/imperialdsi>

1 0

Departmental Computing Meeting this Friday
by David Colling 08 Jun '16

08 Jun '16

Dear All, The Computing meeting will be in 1004 at 12.30 on Friday. Exam marking this week has meant that I haven't given the agenda as much thought as it deserves but now that I am sitting in an airport lounge below is a possible agenda prototype (I will work on it tomorrow but this is to get people thinking and ideally provide feedback): - News from ICT (Person decided by ICT) - Status of the desktop/laptop replacement scheme (Harry V.) - UG computing: o Software to installed on the teaching machines o Unified distribution of python across college. ICT have suggested providing our own in the past but this proved hard. o The college condor batch system for desktops and how we might use it o the Physics data disks and how we might use them very soon. - Teaching IT facilities in general (traditionally everything from projectors to services available) - PG/RA teaching/development: o what can we do help our PGs and RAs in continuing to develop their software/computing skills? o Do we want to offer advanced programming courses? Perhaps cross section or perhaps across the whole department? If so in what languages? o Computing/Software fellowships. o Other forms of career development. - Research Computing: o On what facilities do different research groups rely? (it is a while since we have discussed this and suspect that it has evolved) o What facilities would make a difference (in a positive way). I suspect that there is little that we can other than encourage on this one. o Other concerns from the groups... - The national picture for a UK RCUK (RUK) infrasture... what is happening and who knows what about this (pooling knowledge is always good). Different research councils are putting out different calls/ideas. What can we do about them? - Next meeting date. - AoB Looking at it his already looks like a half day meeting, but lets try to keep it sharp and arrange further discussions where required. Any thoughts/pointing out of omissions et c. et c. are very welcome Best, david

1 0