April 28, 2005

middleware (cont)

2 Comments

Good point by Thom Law­rence in the com­ments of the pre­vious post:

This is, more or less, the cen­tral theme of all soft­ware deve­lop­ment best-practice since time began. Pro­ce­du­res, Modu­les, Objects, Com­po­nents, Ser­vi­ces… Mac­Guf­fins of all sorts have yet to resem­ble the sil­ver bullet you’re after. All of these tools are just to defend against uncer­tainty and change.
But it’s not really a pro­duct you can sell. You can’t have your cake made out of com­ple­tely incom­pa­ti­ble ingre­dients, sprin­kle on magic midd­le­ware dust and eat it.
Infras­truc­ture has to be a first class citi­zen in your busi­ness, right from the start. Peo­ple _know_ how to build these sys­tems. Peo­ple are already buil­ding them. It’s just a mat­ter of com­mit­ment and sta­mina.
Tech­no­lo­gists ver­sus cul­tu­ra­lists again?

A while back I wrote a thing about tech­no­lo­gists vs cul­tu­ra­lists. So I sup­pose you can’t build “Skunk­works” soft­ware and hope to sell it at a pro­fit, unless you already have a “Skunk­works” cul­ture ready and willing to use it.
[NOTE TO SELF:] Greed kills skunks. There can be no tech­no­lo­gi­cal solu­tion without a cul­tu­ral solu­tion. Cul­tu­ral solu­tions are more valua­ble and pro­fi­ta­ble than tech­no­lo­gi­cal solu­tions.
[NOTE TO SELF:] Stick to car­too­ning. You are so out of your league.

2 Responses to “middleware (cont)”

  1. john Allsopp says:
    April 28, 2005 at 9:48 pm

    Geof­frey Moo­res tech­no­logy mar­ke­ting clas­sic “cros­sing the chasm” addres­ses this very well. As far as I know, the ori­gi­nal Skunk­works deve­lo­ped mili­tary airc­raft. Did any com­mer­cial airc­raft come from that? Perhaps indi­rectly. I men­tion this because the mili­tary are in a sense your clas­sic “early adop­ter” who are willing to make tech­no­logy work at just about any cost to get a com­pe­ti­tive advan­tage. Geez they are willing to make ham­mers and toi­lets work at just about any cost :-)
    A skunk­works approach will defi­ni­tely get you star­ted, but it won’t get you into the mains­tream mar­ket. For that you need to deve­lop a “whole pro­duct”.
    Any­way, for those with tech­noogy product’s, Moo­res books really are a must read.

  2. Ric says:
    April 29, 2005 at 12:33 pm

    john: Skunk­works won’t get you into the mains­tream, but is that where you want to be? That’s where you are figh­ting against com­mo­dity soft­ware, entrenched (and ossi­fied) mana­ge­ment, and zero smart con­ver­sa­tions. Infor­mal net­works, both inside and outside the orga­ni­sa­tion, are already where most of the ‘good’ (sorry — value jud­ge­ment there) work gets done even now, and ‘mains­tream’ soft­ware, just like mains­tream mana­ge­ment, is about doing the same things with con­sis­tency, effi­ciency and repea­ta­bi­lity. Now for SOME stuff in the busi­ness, that’s exactly what you need (nobody wants the pay­roll sys­tem to work like a lottery!) — but growth and inno­va­tion and reten­tion of the smart peo­ple is going to need something a little more adven­tu­rous than that. I’ve star­ted thin­king about some of this (I’m a soft­ware mana­ger caught bet­ween a rock (ERP imple­men­ta­tion) and a hard place (the “orches­tra­tion” some­body men­tio­ned in a com­ment on the pre­vious post)) — http://aqualung.typepad.com/aqualung/2005/04/to_dev_or_not_t.html
    and I want to think a lot more about it, because I can see my own job disap­pea­ring into the gap bet­ween my per­so­nal phi­lo­sophy of how soft­ware sup­ports a busi­ness and how the busi­ness is imple­men­ting software.