Will we get a $5000 custom app development business?

There are a lot of apps I want that that I can't find, apps like a personal-corporate search tool, or a corrosion-resistant wiki, or adding a graph navigation (link) layer to a plain text repository.

I know enough about building software that I can, fairly quickly, make up an initial set of requirements that I'm reasonably sure are a good match to readily available toolkits and technologies. I can quickly revise them to match timelines and resources. It's a bit of an odd knack but it's one I have.

There are architects I know who can, in an hour, lay out how to build to those requirements -- including specifying platform options and toolkit alternatives.

The first step is free for me. For $1K-$2K for 1-2 hours work I can find good people to do the second step.

Things get tricky when we try to turn this into code. Anyone who has done outsourcing knows that there's a gap between the theory of Ugandan MOOC grads wanting to work for $20/hour and the reality of high school quality coders who work for a year before moving into management. Not to mention little details like testing.

Open source only works if the app is something the developers themselves want to use ... and that's a special case.

So for $2K I can only get to the pre-coding step given today's methods for organizing paid work ...

The corporate wiki I want

The modern publicly traded corporation is to data as water is to iron. It is a challenging environment for use of anything but the most corrosion-resistant materials: .doc, .ppt and .xls. It's not a good home for things like Wikis.

So we need a corrosion-resistant Wiki. This is what I think we need:

• Minimal dependencies, maximal configuration simplicity. This probably means file based, no database dependencies.
• No expertise for setup and operation - if you can use basic Word and files you can implement and move the Wiki in under 1 hour.
• Rapid and perfect portability. Rapid moves, no broken links.
• WYSIWYG (roughly) and a native syntax that works. Probably markdown.
• Strong search API
• Cross-platform (shudder. Does this mean Java for the environment?)
• Low cost or open source/free. (Freemium model ok for scale)
• Maximal link preservation (Confluence, I can't believe how readily you break external links. This is simple stuff.)

There are probably other requirements, but I only have five minutes. Those are a good start.

 

This wiki doesn't exist today, but I am certain it could be built. The next time I'm unemployed it's on my list (along with a lot of other things, alas)

See also: 

• Wiki impressions: XWiki

Evolutionary economics is due for a reboot

Jupiter's red spot is a transient thing. One day the storm will dissipate -- though probably not in our lifetimes.

Life is a transient thing; a deviation from general entropy flows. One day life will go away.

NFL subsidies are a kind of emergent trap, one day they will go away.

Our universe, lives and economies are rife with transient exceptions. In the long run they go away, but we don't live in the long run. In our time frames the exceptions are the rules.

These exceptions don't appear in micro-economics, macro-economics, or even behavioral economics. Macro tells us average temperature, it doesn't describe the tornado bearing down on our house. 

So we need something radically new. Something like evo-Econ or eco-Econ or Canopy Economics; a discipline that focuses on self-organizing emergent phenomena, that recognizes that natural selection is inevitable in a complex adaptive system.

Did I put enough buzz words in there? Because I'm not dumb enough to think these are new ideas. Ecological economics is the application of economics to ecological topics, so not what I'm thinking of, but Thorsten Veblen coined the term Evolutionary Economics in 1898. Judging from the wikipedia article Evo-Econ has veered off in several different directions over the past 110 or so years -- but sometimes it tracks closely to what I'm describing here.

It's due for a renaissance today.

Project Memfail: Tackling my search space problem

I've hit the Wall.

It's partly entropy-related wetware failure, but I'd be in trouble even if I were immortal. I have two many search-spaces and information stores in too many places.

Things aren't so bad in my personal domain - I have two search spaces. My Simplenote files (via NvAlt), Email (Mail.app) and Google Docs (via Google Drive and CloudPull [1])  are mirrored back to my Spotlight search space, and I use a Google custom search engine against my blogs, archived web site and app.net streams. So my two personal search spaces are private/secure and public. I can manage that [2], and I'm careful not to add anything that would require a third search space.

In my work domain though it's a mess. I have information scattered across several Wikis, multiple document stores, my local file system and multi-GB email store, and the remnants of blog whose server died. All of this in an environment that, for multiple reasons, is driving to an information half-life of 1 year. I have too many search spaces; I can no longer track them all.

So I'm launching Project Memfail :-). I need to rescope my search spaces - esp. my corporate one.

- fn -

[1] My main information loss over the past decade was GR Shares, I have some recovery there via CloudPull.

[2] Native spotlight search has issues, but I can work around those. Of course when Google Custom Search dies I'll be in for a painful transition.

New age data loss: what do we do when backup isn't enough?

A few weeks ago my primary drive began to fail. This was painful because the recovery process exposed lots of rusty plumbing, but I was never worried. I have onsite and offsite backups, multiple iterations of each, using two different technologies (Time Capsule and Carbon Copy Cloner).

I feel pretty good about data loss due to drive failures or boneheaded mistakes.

I don't feel good about the iTunes 11.1 making my media disappear. If not for a chance post encounter I probably wouldn't have discovered the loss for months -- at which time recovery might have been impossible. In this particular case the media files weren't lost -- but iTunes 11.1 couldn't recognize media kind metadata assigned by an earlier version of the app.

Similarly Aperture and iPhoto have been known to lose track of video and images; both have added features to look for orphaned files thanks to past experiences. Users have to know to run these procedures however.

Even worse are sync errors. Apple's Discussions forum are rife with reports of data loss or corruption related to iCloud use. This is bad when it's obvious, but far worse when it goes undetected.

We need new approaches.

We need a utility that keeps a record of deletions, and has rules to notify us of unexpected deletions. That's doable, I'm looking forward to buying a copy.

I don't know what to do about data invisibility arising from application database corruption or bugs like iTunes media kind metadata conversion failure. That's a lot more subtle. Given Apple's poor record of managing these problems (I can think of several things they could do) I wonder if they'll need someday to be legally liable for gross negligence leading to data loss. In the meantime, I suppose we could Voodoo stick pins in something -- or rant in our blogs, which is probably about as effective.

Facebook's real problem -- people sharing photos and stories

Andrew Leonard writes about Facebook's awful ad targeting, and their efforts to retarget ads and posts. That made me wonder why Facebook is determined to drive away me, my friends and family, and my neighbors.

It's either incompetence or design. I suspect the latter. Facebook doesn't make any money off us now, and I suspect they gave up hope when their farm games died.

Facebook wants a passive audience consuming celebrity endorsements and product placements. The TV audience of old, or the Twitter audience of today.

People sharing stories and photos are a distracting waste of bandwidth.

Amerisclerosis? Why US underemployment (and probably inequality) persists.

Brad DeLong excerpts Olivier Coibion, Yuriy Gorodnichenko and Dmitri Koustas: Amerisclerosis? The Puzzle of Rising U.S. Unemployment Persistence: "The results suggest that only cultural factors can account for the rising persistence of unemployment in the U.S., but the evolution in mobility and demographics over time should have more than offset the effects of culture."

I can't tell from the excerpt what is meant by cultural factors. It could be "declining labor mobility, changing age structures, and ... decline in trust", in which case the term has a technical meaning rather than the usual way we think of culture.

The "decline in trust" is weird for a non-specialist, I assume it too has a technical meaning.

Declining labor mobility is a bit of a black box; for example if automation were concentrating employment opportunities in high EQ/IQ positions labor mobility would be more genetic than cultural.

I continue to follow these discussions from a non-specialist perspective. I don't know how this can be tested in economic frameworks, but I continue to think we live in a world where the rate of change, largely driven by globalization (India/China) and information technology, has outstripped our ability to adjust and adapt. Perhaps we'll catch up if/when change stabilizes.

Beyond that, I think the rise of massively powerful corporations that dedicate wealth to legal and financial manipulation rather than innovation, combined with an increasingly aged (and, in the true sense of the word, conservative) electorate, is worsening America's ability to adapt.

Mounting a bike light when you use a front bag

My NiteRider Lumina 700 seems solidly made, but since I use a front bag it's a nuisance to mount. I took apart the helmet mount and fit that into an old reflector mount that puts the light in a good position. I think this mount will work, but in case it doesn't I put together a list of alternatives. (Of course the ideal front bag would include a bike mount, but mine doesn't. I prefer not to use the helmet mount, but that can work too.)

Most of these mounting solutions are designed to work with a handlebar clamp mount like the one that comes with the Lumina:

• Gino Light Mount from Paul Component Engineering $24, uses Frame eyelet.
• Sheldonbrown.com has several mounting solutions that might be useful (Sheldon died in 2008, but Harris Cyclery lives on)
• Lamp Holders : • Milwaukee Bicycle Co. • Ben's Cycle: lots of options here, many designed for bicycle computers but will be fine with this light.
• Front Light Brackets products from SJS Cycles: UK, but lots of ideas here
• Cronometro Nob: I like this design for a fork mount.
• Origin8 Light Mount Frame/Fork Eyelet Stub B $10
• Problem Solvers Quick Release Light Mount: replaces nutted side of front QR Skewer to provide mount
• Problem Solvers brake stud mount
• Urban Velo - DIY rack or fender mount

I ended up ordering a Paul Gino Light Mount from Amazon for $17. I saw a similar "Origin8 Light Mount Frame/Fork Eyelet Stub B" on Amazon, but Origin8's web site doesn't have a light mount and the reviews were inconsistent. I suspect this might be a counterfeit based on the Gino.

Of course another option is to go all in and buy a front rack, then mount the lights there.

Update 9/30/13: I bought the Paul Gino. It seemed small at first, but of course handlebar diameter isn't big either. It's quite finally made. I expected the M5 mount screw would fit my touring bike's front fork carrier mounts, but first I tried it on an old reflector mount that came with the bike. That mount, with a jury-rigged platform, had held my older front light, but my Nite Rider Lumia 700 was too heavy for it.

As shown below, it worked quite well. The reflector mount, as it turned out, is M5 threaded and I put a solid torque into the steel bolt and steel mount. The Lumia 500, set a few cms to the right of center, just barely cleared my cantilever brake cables.

Perfect result! 

BTW, many people like to mount lights on the lower to mid fork to get a better view of the road surface. I like to see the road too, but my main goal is to be seen by cars. So I prefer this mount - visible to cars, but a good view of the road.

BBS Dial up information from 1980s

I came across a list of the BBS sites I used to get to via pre-internet call routing. I think I used something like Telnet. The NSA probably wasn't monitoring our traffic.

1- The Crow's Nest 786-3161 38400 N81 F 0
2- Beowulf's Club 21 789-0233 38400 N81 F 0 BEO
3- Eagle's Nest BBS 789-1792 38400 N81 F 0 EAGLE
4- Mich Tech PC UG 1-487-2738 38400 N81 F 0
5- Lighthouse BBS A,428-3425 38400 N81 F 0 LITEHOUS
6- UPRNet BBS A,786-2179 38400 N81 F 0
7- ICU BBS D,428-3250 38400 N81 F 0 ICU
8- The Cardboard Box E,428-9135 38400 N81 F 0 CBB
9- PC exchange (lol) 13138490499 38400 N81 F 0
10- PIL Software Systems 13144499401 38400 N81 F 0
11- procomm user support 13144749543 38400 N81 F 0 PRCM
12- PC Resource BBS 16039249337 38400 N81 F 0
13- western union 18003254112 38400 E71 F 0
14- MIRNET ,,1,,786-2179 38400 E81 F 0
15- Sparta Board 1-201-729-7056 38400 N81 F 0
16- Shareware Dist Network 1-203-634-0370 38400 N81 F 0
17- NASA Space Link BBS 1-205-895-0028 38400 N81 F 0
18- Invention Factory 1-212-431-1194 38400 N81 F 0
19- NYPC BBS (Steinberg) 1-212-633-1870 38400 N81 F 0 NYPC
20- Magna Carta Software 1-214-226-8088 38400 N81 F 0
21- Free Net (Med BBS) 1-216-368-3888 38400 N81 F 0
22- Genie 1-301-340-5565 38400 N81 F 0 genie
23- Qualitas BBS 1-301-907-8030 38400 N81 F 0
24- NIST Comp Sec BBS 1-301-948-5717 38400 N81 F 0
25- Medical Sftwr XCH BBS 1-305-325-8709 38400 N81 F 0
26- ComputerDirect BBS 1-312-382-3270 38400 N81 F 0
27- IBM BBS (Atlanta) 1-404-835-6600 38400 N81 F 0
28- Pitt XPress Support Serv 1-412-864-2294 38400 N81 F 0
29- PKWare Support BBS 1-414-354-8670 38400 N81 F 0
30- The Home Computer BBS 1-414-543-8929 38400 N81 F 0
31- xerox bbs 1-414-797-5216 38400 N81 F 0
32- the Well (Whole Earth) 1-415-332-6106 38400 E71 F 0
33- ATI BBS (Video) 1-416-756-4591 38400 N81 F 0
34- The Softstone 1-502-241-4109 38400 N81 F 0
35- Micro Cornucopia BBS 1-503-382-7643 38400 N81 F 0 mcr-corn
36- Intel Tech Support BBS 1-503-645-6275 38400 N81 F 0
37- Central Pt. Software PCT 1-503-690-6650 38400 N81 F 0 cps
38- BMUG (Berkeley) 1-510-849-2684 38400 N81 F 0
39- The List: PD SW BBS 1-516-938-6722 38400 N81 F 0
40- InterNet (EMC2) 1-517-353-8500 38400 E71 F 0
41- SoftLogic BBS 1-603-644-5556 38400 N81 F 0
42- Dr. Dobb's BBS 1-603-882-1599 38400 N81 F 0
43- SSI 1-609-921-7079 38400 N81 F 0
44- Utilities Exchange 1-614-488-3991 38400 N81 F 0
45- NoGate BBS (Pak) 1-616-455-5179 38400 N81 F 0
46- Boston Computer Society 1-617-332-5584 38400 N81 F 0
47- ByteNet 1-617-861-9764 38400 N81 F 0 BYTENET
48- BYTE DemoLink 1-617-861-9767 38400 N81 F 0
49- Salemi Doghouse BBS 1-703-548-7849 38400 N81 F 0
50- Government (IRS) BXR BBS 1-703-756-6109 38400 N81 F 0
51- Online Soft. Srch. 1-704-255-8259 38400 N81 F 0
52- Western Digital Tech Sup 1-714-756-8176 38400 N81 F 0
53- Gibson Research BBS 1-714-830-3300 38400 N81 F 0
54- FDA BBS 1-800-222-0185 38400 E71 F 0
55- CSV Phone List 1-800-346-3247 38400 E71 F 0 csv_phn
56- MCI FAST 1-800-456-6245 38400 N81 F 0
57- Cristal 1-800-527-0531 38400 E71 F 0
58- PDQ (NCI) 1-800-546-1000 38400 E71 F 0
59- ETNet 1-800-546-1000 38400 E71 F 0
60- U of M Library 1-800-669-8779 38400 E71 F 0
61- Thous. Oaks Tech. Exch. 1-805-493-1495 38400 N81 F 0
62- The Shadowland --Chris M 1-814-238-4654 38400 N81 F 0
63- Internet - Caltech 1-818-405-0161 38400 N81 F 0
64- Quarterdeck BBS D1-213-396-3904 38400 N81 F 0
65- Trident bbs A,1-415-691-1016 38400 N81 F 0
66- Micro-Sellar D,1-201-239-1346 38400 N81 F 0 MSELL
67- OS/2: Fernwood D,1-203-483-0348 38400 N81 F 0
68- GMED BBS II D,1-301-402-7857 38400 N81 F 0
69- Symantec BBS D,1-408-973-9856 38400 N81 F 0 SYMANTEC
70- Exec-PC D,1-414-789-4210 38400 N81 F 0
71- OS/2: Shareware D,1-703-385-4325 38400 N81 F 0
72- compuserve D,1-800-848-4480 38400 E71 F 0 CSV
73- Sound-Advice BBS D,1-816-436-5635 38400 N81 F 0
74- Always BBS D,1-818-597-0275 38400 N81 F 0 ALWAYS
75- MCI Mail E,1-800-333-1818 38400 N81 F 0 MCI
76- Grateful Med BBS F,1-800-525-5756 2400 N81 F 0 GMED