|
Mittwoch, 24. Juli 2002
Tobi Schäfer in seiner Eigenschaft als Inhaltsverwaltungs-Rechercheur: "for small websites which do not need heavy (pseudo) content management features at all (which is the case for the project i did the current research for), i would rather build a custom system from scratch with a scripting language of my choice than trying to learn another concept superstructure for achieving the same in twice the time." — How right. The sweet spot between useful substrate and superstructure overcomplexification is a tough one to find ...Der Flash Communication Server scheint zu halten, was Macromedia versprochen hat. Eine minimale Chat-Applikation ist problemlos in ~ 15 Zeilen möglich und auch sonst scheint damit alles, was das kollaborative Herz je begehrt haben mochte einfacher geworden zu sein als je zuvor.Warren Buffet: "Who Really Cooks the Books?" — "In calculating the pension costs that directly affect their earnings, companies in the Standard & Poor's index of 500 stocks are today using assumptions about investment return rates that go as high as 11 percent. The rate chosen is important: in many cases, an upward change of a single percentage point will increase the annual earnings a company reports by more than $100 million."
| earl 8141 days AGO wie wahr, wie wahr - das selbe gilt wahrscheinlich bei warenwirtschaftssystemen fuer KMUs. und nicht nur der break-even point zwischen sinnvollem substrat und absolutem overkill ist interessant - auch moeglichst universelles substrat zu basteln, erscheint mir interessant. |
| David Ness 8141 days AGO I'll add my `How true' to the list as well. This is a point all too often neglected. Perhaps part of the explanation is that having all coded something too specifically, we are then prone to swing to the other extreme and overgeneralize. At Wharton we used to talk about middle out development strategies just because being `closer to the problem' was being closer to the sweet spot. |
| stephan schmidt 8141 days AGO How true.
I manage my homepage now with a wiki, which does all I want and fast (after an EJB 500 classes nightmare ;-). |
| tobi 8141 days AGO i might add something here, too: it's always very easy to define and find the, erm, content management system (or call it publishing system) of choice for yourself. but as usual the biggest problem when working for a client is the client itself (to put it nicer: you never know if the client can cope with a piece of software as you can do).
currently, i am having a hard time to make a decision for a low-cost system that indeed is usable for the client. either it is simple but inflexible (ie. almost impossible to extend for custom functionalities) or it is full of possibilites i just cannot use because i run into the superstructure problem (which is also easier if you want to learn it for yourself, for fun).
that's why i came to the conclusion that doing it from scratch sounds like the best way. we'll see, though (client has to nod before any action is taken, anyway).
and btw. it's very likely that i cannot use helma in this case because of the server capabilities. although i am already thinking about a helma app that publishes static html from a local machine via ftp... |
| stephan schmidt 8141 days AGO We wrote SnipSnap from scratch because we can control what features are in and what features are oversized, e.g. Zope is nice and fine, but a lot of features in Zope are not needed.
We want to introduce the wiki as a CMS here at Fraunhofer and run into the same problems with clients. Most often they want features they will never use. But it's kind of nice to have them. This leads to generalization (which is typical for computer scientiests I guess).
But the vanilla macro idea (which we copied but modified) is very powerful and we try to solve a lot of problems ("I have read this meeting announcement") with macros to stay lightweigt.
Has someone expirience with introducing lightweight CMS solutions into organizations ? |
| andi 8141 days AGO yes i do have... in the last company i worked for, a guy coded a small article publishing system, which could even generate some stock charts, in only two days, with the clients projectmanager sitting next to him. it worked really well for two years and produced tons of static content for a part of the site and was definately worth the money we charged. on the other hand, doing two redesigns was a nightmare when it came to that application and in the end, when we put the whole site into a bigger CMS, it took us three days with SED and AWK to migrate the pages produced by the smaller system. if i had to do that again, i would try to convince the customer to invest a little bit more into the future. |
| earl 8141 days AGO somewhat related: the "European Workshop on Industrial Computer Systems Reliability, Safety and Security" (www.ewics.org/) uses Vanilla to manage their site - they use a quite old version (0.3.6) and seem to be happy with it for more than two years now :) i've been in no way personally involved with this however - maybe chris has a bit more to tell :) |
| chris 8141 days AGO For me, Wiki-like things (obviously ;-) work very well too. About the only thing that I really miss in Vanilla (when compared to the Hop or other "real" application-serveresque things) is the speed of in-memory data shuffling ... CGI just doesn't cut it in terms of scalability. For a few mid-traffic sites on a single server this isn't a problem (thanks to modern hardware), but the amazing level of Hop scalability as showed off by Antville.org (~ 2000 sites hosted on a single server without serious performance problems) is definitely out of reach.
But well, this is merely an implementation detail ;-) |
| chris 8141 days AGO In response to Stephan, I think the art is to convince "customers" (internal or external) to let structures emerge, i.e. most often a simple text item with event items is enough, but they think they want a full-fledged data-base-backed event management solution so that they can sort their 5 events by date, time, alphabetically and by popularity. |
| earl 8141 days AGO hehe - so true! let structures grow naturally and communities evolve. too much artificially imposed structure kills |
| gavin 8141 days AGO The problem with biologic metaphors in the computing world is that there're no automatic means of killing things off, i.e. resources run out, organism dies. I've worked with a large number of home grown CMS systems that really deserve to die. |
| stephan schmidt 8140 days AGO Many thanks to all suggestions.
I'm amazed by antville, too. I wonder whether antivlle runs one DB that is mutli-weblog aware (a weblog name tag to all tables) or different DBs (the tables unaware of other weblogs). |
| chris 8140 days AGO One DB. Performance dank dem Object Cache helmatique. |
| andi 8140 days AGO by the way, is there a reason why antville uses FESI (js1.3) instead of Rhino (js1.5)? i know thats not the right place to ask but i am too lazy to get an account ;-) |
| chris 8137 days AGO Yeah. The historical reason ;-) |
Please log in (you may want to register first) to post comments!
|
|
SEARCH
GET YOUR MOVE ON
ALMOST ALL ABOUT YOU
So log in, fella — or finally get your langreiter.com account. You always wanted one.
Nearby in the temporal dimension:
Nobody.
... and 152 of the anonymous kind.
Click on for a moderate dose of lcom-talk. This will probably not work in Lynx and other browser exotica.
THIS DAY IN HISTORY
2005 / 2004 / 2002
BACKLINKS
none
RECENT EDITS (MORE)
films-seen Blood Stone y!kes wet towel B Studio Pilcrow News Nastassja Kinski 2011-10-06-steve 2011-10-06 comment-2011-08-04-1
POWERED BY
&c.
GeoURL
RSS 0.92
FRIENDLY SHOPS
Uncut Games bei Gameware
OFFEN!
Offenlegung gem. §25 MedienG:
Christian Langreiter, Langkampfen
See also: Privacy policy.
|
|