-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am 06.11.2011 18:11, schrieb Colin Yates: > I suspect programming in clojure effectively is as much about > changing the *way* you develop as much as it is about about the > shape of the solution. It sounds as if you are suggesting this is > so.
yes. you are (almost) never going to write the final code. you are going to write code which you are going to change later, so code should be written in a way that allows you to do so quickly. > I am about to start my first clojure project tomorrow which is to > replace an existing web app which is a case study in how short > term wins quickly build up technical and architectural debt. you *can* have short term wins without adding complexity. you just can't have it always ;) It is also a > wonderful example of how painful it is to live in a world where > the code base is so bad that changing it is too dangerous so lets > add another layer ..... Yuck. > > How did you find IntelliJ iDEA for clojure? very stable, but aside from the project/module setup, it's barely more than a text editor with symbol highlighting, basic code completion/auto import and basic function call resolution. compared to the groovy, scala or even java capabilities of intellij - no contest. i guess given the dynamic nature of clojure, you'd have to put a LOT of resources in developing useful code analysis algorithms. I have used both it and > eclipse (STs) for years but am considering emacs (with which I have > a beginner's competency). I am intentionally not using spring or > hibernate and I expect the project will be fairly small in the > short term actually so I think emacs might be achievable.... (I > am also switching from spring MVC and jquery to conjure and yui3 - > this is all about using the right technology from the start and I > love vert iCal learning curves :) > > To explain why this madness might actually be a great idea - the > existing app will be deployed with some additional json services > so the new app will initially just be a very thin skin that uses > those json services in the short term. Later on the functionality > will be rewritten in the new app. > > Exciting times! > > Sent from my iPad > > On 6 Nov 2011, at 16:49, Dennis Haupt <d.haup...@googlemail.com> > wrote: > > "special cases" that depend on mutable state are evil. i avoid > mutable states as much as possible, no matter which language i am > using. what i meant were cases where i roughly think about 3/4 of > all cases, start coding and along the way i notice that in one case > in need another input parameter for my function and have to > refacto everything i have written so far. > > my experience is that you need to balance out planning and actual > coding. there's a limit for planning. if you plan too much, it > won't help. you'll just introduce problems and won't notice it > because you are skipping the reality check. reality never misses a > case, but even the smartest guy/gal does :) > > what i do in <random oo language> is to just start somewhere and > code recursively. as soon as something gets to complex, i split it > up or do minor refactorings. this works pretty good and i rarely > end up in a situation where i have to go back a big step. i rarely > lose speed, even if i encounter a surprise. > > i'll become faster once i get used to clojure, but i doubt i'll be > able to be as fast as in scala or java. this is not because the > language itself is bad. java is really verbose, but that doesn't > matter at all. i don't write java code. i use intellij idea. it > writes and rewrites java code for me. with equal tool support, i'd > prefer clojure over java. > > > > > > Am 06.11.2011 14:48, schrieb Colin Yates: >>>> But aren't the "edge cases" fewer given the notion that >>>> functions should be entirely defined by their inputs as >>>> oppose to being dependant on mutable state external to he >>>> function, in the most part. >>>> >>>> I am agreeing with you, and find these real world >>>> experiences incredibly useful. >>>> >>>> Sent from my iPad >>>> >>>> On 6 Nov 2011, at 13:03, Dennis Haupt >>>> <d.haup...@googlemail.com> wrote: >>>> >>>> Am 06.11.2011 13:44, schrieb Colin Yates: >>>>>>> I have a hunch (as oppose to IME as I don't have any >>>>>>> yet :)) that Clojure doesn't require as much due to at >>>>>>> least the following points: >>>>>>> >>>>>>> - there is much less incidental complexity and >>>>>>> ceremony required to manage (and therefore refactor) - >>>>>>> implementations in Clojure require more thought and are >>>>>>> much more focused on one thing then in Java, therefore >>>>>>> there is a much higher chance that you get it right the >>>>>>> first time >>>> >>>> this is a double edged sword. you *do* get it right *if* you >>>> think it through, but reality is often more complex than you >>>> assume. if you suddenly see that you forgot to handle special >>>> case x, you are punished harder than in <random statically >>>> typed oo language>. >>>> >>>> in oo, you can do a few "emergency hacks" before everything >>>> becomes unmaintainable. in clojure, i could not do this >>>> without committing maintenance suicide immediately. for each >>>> case that popped up surprisingly, i had to do a bit of >>>> refactoring. >>>> >>>> that whole "dynamically typed" thing is creeping me out. >>>> >>>> >>>>>>> >>>>>>> To put it another way, how much refactoring we we do in >>>>>>> Java is down to managing the complexity of the problems >>>>>>> inherent in Java the language; complecting state and >>>>>>> identity for example. >>>>>>> >>>>>>> A concrete example: I know I often refactor code out of >>>>>>> one class into a utility once that becomes useful by >>>>>>> other classes. The reason I need to do this is because >>>>>>> in java the functionality is ring fenced by the data >>>>>>> that it works on (I.e. encapsulation in the container >>>>>>> called a class). In clojure that association can be >>>>>>> made but right off the get-go data and functionality >>>>>>> are separate. >>>>>>> >>>>>>> Just my two-pence worth and I still haven't written >>>>>>> more than 2 lines of clojure so it probably isn't worth >>>>>>> 2p :) >>>>>>> >>>>>>> Sent from my iPad >>>>>>> >>>>>>> On 5 Nov 2011, at 12:16, Dennis Haupt >>>>>>> <d.haup...@googlemail.com> wrote: >>>>>>> >>>>>>> hi, >>>>>>> >>>>>>> i'm half done with my asteroids clone. i stumbled over >>>>>>> a few problems and wanted to know how others already >>>>>>> solved them :) >>>>>>> >>>>>>> i am used to "less concrete programming". i ask my >>>>>>> tools to do the actual analysis and coding for me: * >>>>>>> where is that used? * rename that * show me all errors >>>>>>> * add a parameter here * autocomplete this for me, i am >>>>>>> too lazy to even remember the exact method name * show >>>>>>> me a list of all methods and fields starting with an >>>>>>> "e" that fit in here. i know the one i want is in that >>>>>>> list. >>>>>>> >>>>>>> as elegant as clojure may be as a language, it's >>>>>>> dragging me down to almost native text editing which i >>>>>>> did more than 10 years ago when i wrote a game in turbo >>>>>>> pascal. >>>>>>> >>>>>>> how did you solve these problem and get some >>>>>>> tool-support? or don't you miss it because of something >>>>>>> i am not aware of? >>>>>>> >>>>>>> >>>>>>> >>>>>>>> >>>>>>>> -- You received this message because you are >>>>>>>> subscribed to the Google Groups "Clojure" group. To >>>>>>>> post to this group, send email to >>>>>>>> clojure@googlegroups.com Note that posts from new >>>>>>>> members are moderated - please be patient with your >>>>>>>> first post. To unsubscribe from this group, send >>>>>>>> email to clojure+unsubscr...@googlegroups.com For >>>>>>>> more options, visit this group at >>>>>>>> http://groups.google.com/group/clojure?hl=en >>>>>>> >>>> >>>> >>>>> >>>>> -- You received this message because you are subscribed to >>>>> the Google Groups "Clojure" group. To post to this group, >>>>> send email to clojure@googlegroups.com Note that posts from >>>>> new members are moderated - please be patient with your >>>>> first post. To unsubscribe from this group, send email to >>>>> clojure+unsubscr...@googlegroups.com For more options, >>>>> visit this group at >>>>> http://groups.google.com/group/clojure?hl=en >>>> > > >> >> -- You received this message because you are subscribed to the >> Google Groups "Clojure" group. To post to this group, send email >> to clojure@googlegroups.com Note that posts from new members are >> moderated - please be patient with your first post. To >> unsubscribe from this group, send email to >> clojure+unsubscr...@googlegroups.com For more options, visit this >> group at http://groups.google.com/group/clojure?hl=en > - -- -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.14 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQIcBAEBAgAGBQJOtsaYAAoJENRtux+h35aGPG0P/19Hj78/ujhlP3KTxDPhwoKR 1hudW9qqQMbq+XgY4cGVwcgjRuJvUu40zECsqQyHyioAmDLVO+IMcJnPFquogl5O al9RhaIko1s92nxDKTAn6RB4PBhB44hTxlGlLQw5U8okde9FKoKg42h2THB6zhEp DHotPBA28pADNZGHwcayMIxm5GBUfsI415pueneRoLP/R4KudVcDxLj4auGnTMic 702Rj/lqQfIcE8YS8lLmo9hq9yBLy3ELYpzsrfwPvoYhoxsyVxOmO4tmLOETAlGJ yYfG1l51RSH/FLW8PAniyHDIaeYt+JGOyVHtxnok4BJOpHMd1cq6o+lnuKMPUFPb z+0Hf8h5eAixWco9YS0Ues+Th571aSLYfdO92kv+r0aHSxmnUvI5N2AbEihT0nSJ ay2RpQXBMjiWoGcfnANO11nA/VLHyfaZ59w8MaRn2IQ7VhY7IQXcVG/Rza8j/AuT FqvZAd3GwsCx/beiAFC/OXJ/Lo6y30rdaWl5256KLRN+tGzBHYzSlYOHqpFOifnl hdfN6R8gEIK/htm6o5u5x+P0LZSC93VPUaVfqNylAPF1aNhyiN2gJu7nkqOHGoC5 hEQKPzd22rplWsTGTCDp2D0lvR8OXYn6KNiSie/oNY46K6rVyqciGMoWADIHL/zW IlODqaZcm0H6QSlRyZ2A =pMSM -----END PGP SIGNATURE----- -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
