[Twisted-Python] Enterprise, do you copy?
...hello? Hello? Is anyone receiving this transmission? I heard from Dad again today, and he was grumbling about developing in this Oracle JDeveloper environment... Five different languages to develop "java server[servlet?] pages", he says. Java, because that's where Oracle seems to be building its tools these days; then there's Oracle's query language; then HTML for the web interfaces; and JavaScript, because there's something that needs to be done client-side, and, uh, what was the fifth one again? Then there's the hundreds of functions the wizards create for simple tasks, spread across untold numbers of files... Ye gods, man. Can anyone save this tragedy? I can't give you a complete list of requirements, but the general gist is this: The laboratory has researchers and machines, the researchers do experiments and the machines spit out data. Quite a bit of data. This data comes out in raw form, needs to be stored, processed, the results stored. It'd be nice to keep track of which data actually belongs to what run of whose experiment -- it seems like they lose track of that every now and then. Anyway, data (lots, but still in the fractional-terabyte range) stored in Oracle, and, naturally, cross-referenced with other data, and, well, there's a lot of that going around. The researchers, who spend most of their time knocking the brains out of mice instead of clicking on them, need relatively friendly interfaces to the data from their experiments, and for shoveling it around between these various phases of analysis. They hope to add some internet soon, so they can pool data from their distributed colleagues. So, I gotta ask. What's the status of twisted.enterprise these days? With Oracle? Webwidgets would help with the web bit, but they'd still need to write javascript to keep the forms from being completely dumb? Or can you convince them that there's another interface that's portable and *drop-dead-easy* to maintain on the client machines, which he can use as an alternative to web? Is there an IDE which can make up for all the shiny things which the Oracle tools have? If I didn't make the situation sound scary enough before, here's the truly horrifying detail: there's another faction among the computer-guys in the lab which wants to use _exclusively_ Perl. I don't think it's too late. They're still making up their minds. But if there's another option, they've got to know soon, before they're wholly consumed by the dark side. And whatever it is, it's got to do a damn _impressive_ job of selling itself, to overcome both that warm fuzzy feeling imparted by the "Larry Ellison" brand, and the deap zealotry of the Perl encampment. May peace be with you, - Kevin -- The moon is waxing gibbous, 99.9% illuminated, 14.5 days old.
On Thursday, November 1, 2001, at 12:15 AM, Kevin Turner wrote:
...hello? Hello? Is anyone receiving this transmission?
[snip complaint about poorly-connected multiple language environment]
Ye gods, man. Can anyone save this tragedy?
[snip description of big piles of data] [snip description of mouse-blood-thirsty research scientists]
So, I gotta ask. What's the status of twisted.enterprise these days? With Oracle? Webwidgets would help with the web bit, but they'd still need to write javascript to keep the forms from being completely dumb? Or can you convince them that there's another interface that's portable and *drop-dead-easy* to maintain on the client machines, which he can use as an alternative to web? Is there an IDE which can make up for all the shiny things which the Oracle tools have?
I hate to say this, but Twisted is not really in this 'market space' yet. This is a "traditional" (insofar as any tradition has yet developed) web application, and Twisted probably can't help too much. We aim to be there one day, but let me give you a brief description of why we're not yet... Twisted is currently a framework for Python. Not an IDE, not a web-platform development environment. Web accessibility features are an equal partner of several different modes of accessibility. "Enterprise" (relational storage) is still in the prototyping phase, and there are no applications yet that use it. This means it isn't good for quick'n'dirty web apps unless you already know it; it certainly isn't good for quick'n'dirty web apps that you expect to scale. [yet] If you want to have any degree of integrated services (an AIM robot which can be configured through a web page or an IRC server, for example), then Twisted will be an advantage. Also, if you're developing customized client software, PB can be a huge help in getting past the initial milestone of getting your client and server connected. All of these tasks involve a degree of hacking proficiency above and beyond that required to go through a JDeveloper wizard and click on "-YES-! ENTERPRISE! WEB! WOOD GLUE!". I don't intend to slur the intended audience here; I'm saying we would be misrepresenting Twisted to say it's *easy* to add new functionality like that. Of course, the Oracle corporation is lying as well; but that's what they do for a living; I'm just an amateur marketer. ;-) [snip screaming and gnashing of teeth about perl]
I don't think it's too late. They're still making up their minds. But if there's another option, they've got to know soon, before they're wholly consumed by the dark side. And whatever it is, it's got to do a damn _impressive_ job of selling itself, to overcome both that warm fuzzy feeling imparted by the "Larry Ellison" brand, and the deap zealotry of the Perl encampment.
OVERGENERALIZATION MODE ENGAGED I think that your best bet in this particular situation is to befriend the Perl camp. Show them some of the things you can do with Twisted, get them interested in Python. If they want to go with raw perl, they probably understand that having a few expert people generating code can be a lot better than a lot of inexpert people using tools they don't quite understand. They won't be surprised when they hit a rough edge, and won't be shy about contributing changes back to the community. It's likely that you won't be able to convince them, but at least you'll be speaking the same language. Those who are pushing for JDeveloper are beyond hope; we don't have a big name behind Twisted (unless your father's employer knows the true and description-defyingly sinister meaning of "Twisted Matrix Laboratories", which I find unlikely). We don't have big piles of money to wow them with flashy demos, and we don't have any existing "customers" who use the "product" in the way that they expect to. In many ways, the product they are shopping for doesn't do nearly the same thing as Twisted does; they want buttons to click on, not library functionality to use. i-used-to-be-a-java-developer-too,-make-of-that-what-you-will-ly y'rs, -- ______ you are in a maze of twisted little applications, all | |_\ remarkably consistent. | | -- glyph lefkowitz, glyph @ twisted matrix . com |_____| http://www.twistedmatrix.com/
On Thu, Nov 01, 2001 at 11:33:35AM -0600, Glyph Lefkowitz wrote:
[snip description of mouse-blood-thirsty research scientists] [...] We aim to be there one day, but let me give you a brief description of why we're not yet... [snipped]
I think that your best bet in this particular situation is to befriend the Perl camp. [...] Those who are pushing for JDeveloper are beyond hope; [...] All of these tasks involve a degree of hacking proficiency above and beyond that required to go through a JDeveloper wizard and click on "-YES-! ENTERPRISE! WEB! WOOD GLUE!"
Well, let me clarify a little how they arrived at the current position. Dad "retired" from the Oracle arena some years back, when the web hadn't *quite* taken over the world yet. Now he's back in a situation where he has an Oracle database and a need to make interfaces so people can work with this data, and make it otherwise accessable. (Preferrably interfaces which require zero maintance on the client machine.) Of course, this has people saying "Uhh... internet." So he goes to consult the Oracle, and the voice of Larry Ellison speaks to him: "Internet? Enterprise internet web internet. Java!" and Dad comes away thinking, "Oh, so that's how people do Oracle these days." And yeah, there's a strong sense that "creating database interfaces shouldn't require this much *code*, why hasn't anyone else made it as easy as Filemaker has? (too bad there are a host of other disadvantages to developing for Filemaker)", and I know that nothing Twisted currently has is going to be really satisfying in that respect; that's what makes Oracle JDeveloper and its wizardly companions look appealing. But he also knows that in all likelyhood, someone will have to edit some of this wizard-generated code someday, and he's quite concerned that they be able to do so without having them shipped down to the Neurology lab afterwards. If my understanding of the environment is accurate, at least a fair share of the people working on the interfaces are past the "ENTERPRISE! WEB! WOOD GLUE!" stage, so I would think that code maintaniablitity is probably going to be at least as important as clickability. Nobody there currently knows Java, the only argument for it is "it's what Oracle uses today, so if we want to work with Oracle products, this is what we do."
We don't have big piles of money to wow them with flashy demos, and we don't have any existing "customers" who use the "product" in the way that they expect to. [...] "Enterprise" (relational storage) is still in the prototyping phase, and there are no applications yet that use it.
There's a relatively low PHB presence here, so while flashy demos would help, it's not scrictly necessary -- persuasive practical arguments could be effective. However, the second issue -- the lack of maturity and customer base -- is certain to be a killer. Well, I had to ask. Thank you for your honest appraisal. "In-dread-of-being-subjected-to-a-"I'm-learning-java-and-think-it's-stupid"-rant-over-Thankgiving-dinner"-ly yours, Kevin -- The moon is waning gibbous, 99.9% illuminated, 15.0 days old.
participants (2)
-
Glyph Lefkowitz
-
Kevin Turner