The Economist - USA (2022-01-29)

56 Business TheEconomistJanuary29th 2022

ingin(lowcode).Under  the  hood,  this  is

translated into pre­written or automatical­

ly generated code, which then whirs away. 

Such tools are in hot demand. Just 25m

people around the world are fluent in stan­

dard  programming  languages,  reckons

Evans Data Corporation, a research firm—

one for every 125 people in the global work­

force  and  1.4m  fewer  than  needed.  That

shortfall will rise to 4m by 2025, says idc, a

research  firm.  lc/ncproducts  expand  the

pool  of  coders  to  “line­of­business”  em­

ployees who seldom speak c++, Java or Py­

thon.  And  beyond.  Cheryl  Feldman  went

from a junior position in a hair salon to a

technical  career  at  Salesforce,  a  software

firm, thanks to lc/nc. Samit Saini changed

jobs  after  13  years  as  a  security  guard  at

Heathrow  to  become  an  “it solution spe­

cialist” at the airport after making software

on Microsoft’s Power Apps.

Overcoming language barriers

idcreckons  the  low/no  coders  numbered

2.6m globally in 2021. It expects their ranks

to swell by an average of 40% a year until

2025, three times as fast as the total devel­

oper population. The number of organisa­

tions using Power Apps more than doubled

in  2021.  It  now  has  10m  monthly  users.

basf,  a  chemicals  firm,  uses  it  to  let

122,000  workers  write  software.  A  study

last year by Aite­Novarica Group, a consul­

tancy, found that over half of American in­

surers have deployed or plan to deploy lc/

nc. Unqork, a no­code startup valued at ov­

er  $2bn  and  backed  by  Goldman  Sachs,  is

convincing  other  financial  firms  to  take

the plunge. Mr Lamanna envisages a global

population of a billion low/no coders. 

The  dream  of  codelessness  is  not  new.

Tony Wasserman of Carnegie Mellon Uni­

versity’s  branch  in  Silicon  Valley  dates  it

back  to  the  concept  of  “automatic  pro­

gramming” in the 1960s. Since then succes­

sive  waves  of  simplification  and  abstrac­

tion have made life easier for programmers

by  distancing  coding  languages  further

from  the  machine  code  understood  by

computer  hardware.  In  the  early  1990s

Microsoft  tried  to  simplify  things  further

by launching Visual Basic, an early stab at

lc/nc.  In  the  next  decade  firms  like  App­

ian,  Caspio,  Mendix  and  Salesforce  began

offering products aimed expressly at line­

of­business types. 

Recently lc/nc’s potential has been un­

locked by the cloud, which lets people con­

nect  to  data  easily  and  collaborate  in  real

time, says Ryan Ellis, who leads lc/ncpro­

ducts at Salesforce. Last year Amazon Web

Services  (aws),  the  online  giant’s  cloud­

computing arm, introduced Amazon Sage­

Maker Canvas, a set of tools that lets people

deploy  machine­learning  models  without

writing  code.  It  also  offers  Honeycode,  a

no­code app builder, in beta version. 

lc/ncused  to  be  chiefly  about  making

pro  devs  more  efficient.  Now  it  is  also

about  pulling  more  humans  into  creating

applications, says Adam Seligman of aws.

In terms of impact, he says, the latest wave

“will  race  higher  up  the  beach”.  For  one

thing, firms in a hurry to digitise appreci­

ate  that  when  line­of­business  people  de­

sign software, it speeds things up. “A field

worker  making  something  for  other  field

workers is hugely valuable as the feedback

loop is faster,” says Adam Barr, a former Mi­

crosoft pro dev and author of “The Problem

with Software: Why Smart Engineers Write

Bad  Code”.  As  digital  natives  enter  the

workforce  they  are  also  demanding  auto­

mation of repetitive or manual data­entry

tasks, often on pain of walking out. 

In addition, lc/ncis fast becoming the

secret  sauce  in  modern  software  develop­

ment,  notably  in  machine  learning,  says

Arnal Dayaratna of idc. The mastery of Py­

thon or Java required for this type of artifi­

cial­intelligence  (ai)  software  is  daunting

even  for  pro  devs.  Bratin  Saha,  who  over­

sees  aws’s  machine­learning  services,

wants SageMaker Canvas to empower reg­

ular business analysts—from marketing or

finance, say—to deploy machine learning.

That could increase the number of aispe­

cialists available to businesses by an order

of magnitude, he predicts. 

Some scepticism is warranted. Just be­

cause  non­programmers  are  able  to  build

an  application  with  lc/nctools  does  not

mean it will be any good, says Mr Wasser­

man, just as bug­ridden spreadsheets yield

faulty  results.  They  could  also  become  a

headache  for  corporate  itdepartments  if

citizen  developers  collect  customer  data

that  are  worthless  or,  worse,  that  violate

privacy. Especially with no code, business­

es can find that the functionality they need

does  not  yet  exist.  No­code  platforms

make  the  first  90%  of  delivering  a  useful

application easy, and the last 5% often im­

possible, says Tim Bray, a pro dev formerly

ofaws. Andmanyprodevsremainresis­

tant.Althoughtheyturntolc/nctosim­

plifysometasks,plentyofprosseeit asthe

programmingcousinofpin­it­onneckties,

inthewordsofonecommentator.Some

worry that specialising in lc/nc makes

themlooklikedilettantes,reportsMrBarr.

lc/ncwillnotdisplace“full”codingal­

together,asitsevangelistsinsist.Prodevs

willcontinuewritingtheirfirms’corepro­

ductsandmission­criticalenterprisesys­

tems.Buttheywillincreasinglybecomple­

mentedbylegionsofenterprisingline­of­

businessworkerswitha software­develop­

mentstringtotheirbow.Foremployers,

thismeansgreaterproductivity.Forem­

ployees,itcouldbelife­changing.In 2019

theTelstratechnicianbecameseniorbusi­

ness specialistfor fielddigitisation and

hassincebeenpromotedagain.n

Calling all coders

Theshortageeconomy

Morepain,nogain

S

upplychainshaveseldomfeaturedin

companies’ earningsreports overthe

threedecadessinceglobalisationtookoff

inearnest,savefortheoccasionalmention

ofthebenefitsoflowcostsandleaninven­

tories.Thisearningsseason,though,co­

vid­inducedshortagesareamongthefirst

problemsmentionedbymanyfirms.The

Omicronvarianthasworsenedthelogjams

byforcingworkers,inmanyindustriesand

thelogisticsbusinessthatweavesthemto­

gether, to quarantine. And shortages of

bothstaffandmaterialsarecontributingto

inflation,raisingcostsacrosstheboard.

OnJanuary25thdisappointedinvestors

sentge’s sharepricedownby6%afterLar­

ryCulp,theindustrialicon’sboss,saidthat

supply­chain “headwinds” had hit its

Whysupply-chainproblemsaren’t

goingaway

Getting high on short supply

Companies reporting supply-chain shortages

Worldwide survey index

Source:IHSMarkit

10

8

6

4

2

0

21191715131109072005

Outputconstrainedby

materialsshortages

Input items in short supply

1

Long-run average