|Keep It Simple, Stupid|
Re^4: Y.A.N.F: Personal Nodelet enhancements. (criticism)by demerphq (Chancellor)
|on Jun 15, 2004 at 07:53 UTC||Need Help??|
They were about the design being overkill and about a simple text box being better.
Ok, well i think the only staticly held criticism about either sets of patches was a one liner from belg4mit that says the titlebar settings are too complicated. So if the comments you mention were part of discussing that its quite possible I misunderstood them or to be honest completely ignored them on the grounds that the criticism was singularly uninformed and uninformative that I just felt it to be insulting. (Nobody took any time to really review what the TS stuff was actually doing, nor how the features present there would be implemented in a simpler fashion.) So to put it simply had someone written something like this in the wiki or as a reply to the patch then I would have responded directly. So yes its quite possible the comments you are making now occured during a discussion about TS and that i just blanked it out as uninformed comments.
Is that ridicule?
No no. But to me it is a signifigant design difference. Lets consider a couple of subtleties in the PN. Both the original and the relacement are conceived of being a list of link like data, with some subset of N links displayed in the nodelet. So for instance the restriction on a 100 links is iirc in the original code. Now in a freeform text box how do we help the user with this? We cant split the data as its free form. So as they add links the list will grow and grow and grow. Necessating regular visits to cut and paste it down to size etc. Now the code as present has no such problem. Itll only show the number of links the user wants to see (with a hard limit enforced that could be changed if folks want). And itll show either the first or the last N nodes too.
Anyway, unless you can find where someone wrote that down to me in the wiki or something lets not quibble over whether I knew about the simple textbox idea. Just next time be aware that I _do_ and _will_ react to informative constructive criticism. Also be aware that one liners (and im not saying this only to you) like "I dont see the point" and "its too complex" and "I dont want to use it so why should you put it in" are really not what i consider to be either informative nor constructive. At least certainly not as lone comments.
Tell me why your idea is better and you might convince me.
I dont think mixing the idea of a link collector (which is what the PN is in my mind) and a "scratchpad in a nodelet" is a good idea. I think they have different uses, and in some respects are complementary. Above i outline a couple of the reasons. So on that level I dont think there is problem having both. Also, yes my view is that this is very much a "give the lion what he wants for dinner or hear him roar". But that doesnt bother me because i am fairly sure that if youll roar, a bunch of other folks are busy miaowing about the same thing. OTOH, im fairly sure there are lots of folks like me that can see the utility of properly managed list of nodelets and who can also see it being useful to have a free form box as well.
You complain about not getting enough criticism but often don't seem to be able to take it when you do get it
Honestly you and I see certain discussions in different lights. I am always happy to hear useful constructive criticism. However like anyone I dont take catty comments like some of the ones i heard about TS very well. Especially when they come from folks who I haven't seen write any non-trivial patches.
I think the way ive reacted to complaints about my own stuff and other peoples stuff here in the monastery shows that I am responsive to criticism. I had no problem with what the folks said about the NN thing, and i had no problem to putting in the work to extend that to cover as many of the criticisms that I had seen. My response time for responding to nits in my releases, or feature requests or the like should be clear evidence that not only do I accept criticism but i also respond and integrate it as quickly as I reasonably can. But if the thought levels and written criticism that i receive is so shallow that the only permanent record is a comment saying "its too complex" (referring to two seperate patch sets comprising of 1000's of lines of code in probably 10 or so patches) then i don't think im out of line rejecting the criticism as useless.
If those that want to criticise can't take the time to write a thoughtful paragraph or two about what they think is better until months after the original work is done and applied then I think they need to look at their own critical skills and not at my code. I mean if im willing to hack for hour on end on various aspect of the site to make it better then youd think those with strong feelings about how things go could spare the time to write some more useful comments.
As an example I point you to the test server where ysth has the new scratchpad stuff metaphorically spread all over the floor. I had a reveiw of the code and to be honest I think its too complicated for what i think hes trying to do. And I said so, in a fairly detailed note about what my aprehensions were, and also admitting that i dont understand the full design objectives so my criticism could be misplaced. I suspect that ysth has or will read it and come back with some explanations or he'll say that my comments made him rethink and that it is in fact too complicated. Either way he got a lot more than a one liner.
So, to wrap this all up, im happy to take direction in developing the site. Im happy to receive as much constructive friendly well meaning critisim as I can. But im not going to pay any attention to stuff that isnt such. Just as poster to SOPW wont pay any attention to comments to a code review request that say things like "your code sucks". Now when that message comes along with a detailed and useful analysis then its a different story.
Getting back to more constructive thoughts. My current plan wrt to what you want is to provide a way to window the users private scratchpad in a nodelet. Thats an easy proof-of-concept of what you want and can easily be modified to use a $VARS slot instead. It also nicely meshes with existing add to scratchpad functionality so IMO it gets us both soultions for minimal cost. If people really dont see the need for the fancier version once we demonstrate the proof of concept then I will do the work to offline the recent changes.
Thanks for your comments. They are very useful for future work. I will also begin making pre-announcments of my design intentions and plans so that folks can vent their spleens before I put in all the effort to actually provide what they request.
First they ignore you, then they laugh at you, then they fight you, then you win.