Coffee maker

Take a cof­fee mak­er. What is its pur­pose? Its only pur­pose is to make cof­fee. It’s the only thing it does and knows. It has only one but­ton, and every­body know how to use it. Now take the Boe­ing 787. How many but­tons are there? How many peo­ple can land it safe­ly at the air­port with no pre­lim­i­nary train­ing?

All oth­er things be­ing equal or held con­stant, more the de­vice rely on the en­vi­ron­ment and more it does, less is its ease of use. And still, peo­ple be­lieve that soft­ware prod­ucts must mag­i­cal­ly stop fol­low­ing this rule which ap­plies quite well to any real world ob­ject.

Word­Press is a per­fect ex­am­ple. When Michel Valdrighi cre­at­ed the Word­Press pre­de­ces­sor, he wasn't imag­in­ing a sys­tem which will do every­thing you can dream of, in­clud­ing pow­er­ing e-com­merce web­sites. But this blog/CMS plat­form be­came pop­u­lar over time, and users re­quest­ed more and more fea­tures, mak­ing it more and more dif­fi­cult to find the right tool in a huge set of tools for every need.

Re­cent­ly, I was re­quest­ed by a mar­ket­ing com­pa­ny to help their cus­tomer with her Word­Press-based web­site. That was the first clue which showed that some­thing is very wrong: in or­der to just get the things done, the cus­tomer with no tech­ni­cal back­ground needs an as­sis­tance of a de­vel­op­er.

Af­ter study­ing the case, it ap­peared that the sit­u­a­tion was much worse. Here are some is­sues:

• The lay­out of a few pages ap­peared too wrong, and the only way to solve this was to change the CSS. By hand.

• On a few oth­er pages, Word­Press was break­ing the lay­out when sub­mit­ting a page with­out chang­ing it. In fact, since Word­Press re­moves <p/> tags, con­tent which must be sur­round­ed by para­graph tags ap­peared wrong­ly with­out them. By the way, the only way to avoid this be­hav­ior seems to be chang­ing PHP code in ad­min­is­tra­tion files and adding fan­cy tags to the page when edit­ing it. Easy for a per­son with no tech­ni­cal back­ground, isn't it?

• In or­der to mod­i­fy the home page, I was forced to change the cor­re­spond­ing PHP source code. Nice.

• Lots of stuff re­quired more than ba­sic knowl­edge of Word­Press. I try to imag­ine the num­ber of hours I would have spent if I were nev­er us­ing Word­Press be­fore and if I had no tech­ni­cal back­ground what­so­ev­er.

Just with this sim­ple il­lus­tra­tion, it ap­pears that the most pop­u­lar blog/CMS prod­uct fails to be easy to use or in­tu­itive. But it's not sur­pris­ing. When the re­quire­ments grow, the prod­uct be­come more and more com­pli­cate to both de­sign by the de­vel­op­ers and han­dle by the users. The more fea­tures there are, the more ex­pe­ri­enced the end user has to be.

In essence, the end user has to choose:

• Ei­ther he has an easy to use tool which does only few things,

• Or he op­er­ates a few bil­lion lines of code soft­ware prod­uct he must learn to use.

The “pow­er­ful but easy” is not an op­tion here.

What does it mean for the soft­ware de­vel­op­ment mar­ket?

• Pri­mo, it means that we must stop promis­ing pow­er­ful but easy prod­ucts. What hap­pened for years with WYSI­WYG? Com­pa­nies cre­at­ing WYSI­WYG prod­ucts promised that “this time”, they fi­nal­ly achieved a piece of soft­ware which cre­ates high qual­i­ty code but let the user change the con­tent vi­su­al­ly. And the year af­ter that, they said that now, this time, they did it. The sto­ry re­peat­ed at every re­lease of every new ver­sion of WYSI­WYG, con­firm­ing every time that all the mar­ket­ing stuff we heard be­fore was mis­lead­ing. In 2014, I still don't know any WYSI­WYG ed­i­tor which can cre­ate high qual­i­ty code in every sit­u­a­tion.

  In the same way, there is no need to at­tempt to make a mag­i­cal­ly in­tu­itive CMS prod­ucts which can do every­thing you need. We, as de­vel­op­ers, give to our users the im­pres­sion that we are able to give them prod­ucts which will make our work for us. Hope­ful­ly, we con­stant­ly fail while try­ing to do some­thing close to the con­cept. Oth­er­wise, it would mean that we are just a bunch of ro­bots which are paid $100 per hour to do the same work a ma­chine can do for a few cents a month.

• Se­cun­do, by con­stant­ly telling every­one that it's pos­si­ble for a user with no tech­ni­cal skills to cre­ate nice web­sites with out­stand­ing qual­i­ty, per­for­mance, etc., just by us­ing soft­ware as Word­Press, we are sup­port­ing all those in­com­pe­tent pro­gram­mers who will re­peat our words to a cus­tomer vic­tim by telling that for $150, the vic­tim can have a pro­fes­sion­al e-com­merce web­site she could mod­i­fy at any mo­ment. Crap. A pro­fes­sion­al e-com­merce web­site that you can man­age will cost thou­sands of dol­lars, and in some cas­es hun­dreds of thou­sands of dol­lars. For $150, the vic­tim will get a Word­Press core with a bunch of bad­ly writ­ten plu­g­ins down­loaded from in­ter­net, and when the vic­tim would try to make a slight change in the con­tent of her new web­site, ei­ther she will be un­able to find how to change it, or the change will break every­thing all around.

Since this mod­el is es­sen­tial­ly bro­ken, what is the al­ter­na­tive?

For the ba­sic needs, the cus­tomer must in­vest a lit­tle of her time in or­der to learn how things work: learn how to in­stall Word­Press, how to add foot­er and head­er, how to add im­ages, etc. Pay­ing $100 is not an op­tion: the per­son gets the prod­uct, but not the re­quired ex­pe­ri­ence.

For the spe­cif­ic needs or niche prod­ucts, the only choice ap­pears to be hand­made, ex­pen­sive so­lu­tions, in-house IT de­part­ment pro­ject, or some­thing sim­i­lar in­volv­ing tech­ni­cal peo­ple at any stage of the process.