WEBVTT 00:00.000 --> 00:04.640 My mum, Irena, is sitting in the kitchen under the lamp. 00:05.740 --> 00:11.460 In front of her, a new box and that leaflet, folded up like a concertina. 00:12.220 --> 00:18.020 She gets the magnifying glass out, moves it closer, moves it away, 00:19.060 --> 00:21.820 puts a second pair of glasses on over the first, 00:22.740 --> 00:26.320 and then she puts the leaflet down on the worktop and says, 00:26.320 --> 00:30.140 Right, I'll take it the way I always have. 00:31.020 --> 00:35.080 Hello, I'm Tomasz Schuster. I built the Moje Pteczka app. 00:35.840 --> 00:40.680 And what Isa's just described is a moment hardly anybody says out loud. 00:41.280 --> 00:45.560 Because it isn't a story about my mum not understanding something. 00:46.080 --> 00:48.920 No, she'd manage the understanding perfectly well. 00:49.640 --> 00:53.120 It's a story about her physically not being able to see it. 00:53.680 --> 00:55.420 That's not a barrier of knowledge. 00:55.420 --> 00:57.920 That's a barrier of eyesight. 00:58.620 --> 01:00.760 And honestly, I don't blame her. 01:01.580 --> 01:03.840 That print is microscopic. 01:04.760 --> 01:06.880 And here's the first thing I want to say today, 01:07.260 --> 01:09.860 because it settles everything that follows. 01:10.760 --> 01:13.020 A phone reads text. 01:13.820 --> 01:16.040 A phone doesn't read paper. 01:16.560 --> 01:17.080 Hang on. 01:17.740 --> 01:19.360 Then make it read it to her. 01:19.360 --> 01:22.000 I take a photo of the leaflet. 01:22.000 --> 01:23.740 The phone reads it out. 01:24.400 --> 01:25.380 Problem solved. 01:25.760 --> 01:28.440 That's the most common bit of advice people give each other. 01:28.960 --> 01:30.160 And it doesn't work. 01:30.660 --> 01:32.760 What do you mean, it doesn't work? 01:33.160 --> 01:36.740 Because to a phone, a photo of a leaflet isn't text. 01:36.740 --> 01:39.240 It's a picture with letters in it. 01:39.240 --> 01:42.080 But a phone can recognise text in a photo. 01:42.080 --> 01:43.780 I've done that with a receipt. 01:44.300 --> 01:46.200 It can, and that's a fair point. 01:47.040 --> 01:49.300 Newer phones do recognise text in a photo, 01:49.820 --> 01:51.160 and you can lift it out of there. 01:51.880 --> 01:53.920 Only, try it on a leaflet. 01:54.560 --> 01:54.900 Meaning? 01:55.460 --> 01:57.940 Three pages of dense print, two columns, 01:58.520 --> 01:59.580 paper you can see through. 02:00.420 --> 02:02.060 The recognition will lose a chunk, 02:02.500 --> 02:03.460 glue the columns together, 02:04.020 --> 02:04.620 skip a heading. 02:05.420 --> 02:06.980 For a receipt, that'll do. 02:07.720 --> 02:09.280 For something where you want to be certain 02:09.280 --> 02:10.320 nothing's dropped out, 02:10.860 --> 02:11.660 it isn't enough. 02:12.540 --> 02:15.980 So all those read-the-screen features on the phone? 02:16.420 --> 02:17.280 Work brilliantly. 02:17.780 --> 02:20.340 But only on text the phone is handed as text. 02:20.800 --> 02:22.100 Over a photograph of paper, 02:22.540 --> 02:24.020 they simply go quiet. 02:24.560 --> 02:26.100 It's absurd, isn't it? 02:26.520 --> 02:29.620 I've got a device in my pocket that can do anything, 02:30.160 --> 02:33.480 and it won't read Gran a bit of paper out of a box. 02:34.480 --> 02:34.960 Precisely. 02:34.960 --> 02:37.860 And that's why the whole thing comes down to one question. 02:38.380 --> 02:40.700 Where do you get that content as text? 02:42.140 --> 02:43.420 There are four routes. 02:43.960 --> 02:45.260 I'll be straight about each one, 02:45.660 --> 02:47.220 including where it runs out. 02:47.740 --> 02:48.140 Go on. 02:48.720 --> 02:50.220 First, a magnifier. 02:50.740 --> 02:52.700 The camera, the torch, and the zoom. 02:53.100 --> 02:55.760 Together, they make a better magnifier than a glass one, 02:56.020 --> 02:57.620 because it lights the page as well, 02:57.900 --> 02:59.600 and it can push up the contrast. 03:00.040 --> 03:00.940 Mum does that. 03:01.520 --> 03:02.660 Is that wrong, then? 03:03.160 --> 03:04.540 No, it's very good. 03:04.540 --> 03:06.540 Only a magnifier shows. 03:07.020 --> 03:07.840 It doesn't speak. 03:08.380 --> 03:09.300 Both hands are busy. 03:09.700 --> 03:10.740 For checking one thing, 03:11.100 --> 03:12.320 a date on a box, say, 03:12.600 --> 03:13.220 it's perfect. 03:13.860 --> 03:15.540 For three pages of small print, 03:15.800 --> 03:16.540 it's a slog. 03:17.140 --> 03:17.460 Second? 03:18.200 --> 03:19.900 Second, the same leaflet, 03:20.220 --> 03:22.260 but as a file from an official register, 03:22.760 --> 03:24.500 because the leaflet you're holding in your hand 03:24.500 --> 03:27.280 is also published as a document you can download. 03:27.280 --> 03:29.280 So the same thing, 03:29.760 --> 03:30.840 only on a screen? 03:30.840 --> 03:32.300 The same content, 03:32.680 --> 03:34.180 but a completely different medium. 03:34.720 --> 03:36.680 And that's the difference that changes everything, 03:37.160 --> 03:40.040 because a file you can enlarge without it going blurry. 03:40.700 --> 03:42.520 And if there's real text inside it, 03:42.800 --> 03:44.360 the phone can read it aloud. 03:44.840 --> 03:45.300 If. 03:45.300 --> 03:45.380 If. 03:45.960 --> 03:48.200 I don't much like that if. 03:48.660 --> 03:49.580 Quite right, too. 03:50.080 --> 03:51.280 Because some of those documents 03:51.280 --> 03:53.320 are really just pictures of the pages. 03:53.960 --> 03:54.440 Scans. 03:55.060 --> 03:56.920 And then the reader has nothing to read. 03:57.320 --> 03:59.660 And how am I meant to know which is which? 04:00.200 --> 04:01.140 You open the file, 04:01.500 --> 04:02.340 turn on the search, 04:02.680 --> 04:04.960 and type in any word you can see on the screen. 04:05.540 --> 04:06.100 Finds it? 04:06.480 --> 04:07.160 It's text. 04:07.620 --> 04:08.420 Doesn't find it, 04:08.680 --> 04:11.080 and you can't select a single word with your finger. 04:11.600 --> 04:12.360 It's a scan, 04:12.820 --> 04:14.360 and that route's closed to you. 04:14.780 --> 04:15.240 Right. 04:15.640 --> 04:16.980 That's something concrete. 04:17.140 --> 04:17.680 Third route. 04:18.140 --> 04:20.160 The phone reading the screen itself. 04:20.300 --> 04:20.760 Built in. 04:21.060 --> 04:22.520 You don't need an app for it at all. 04:22.860 --> 04:23.760 On an iPhone, 04:23.940 --> 04:25.880 you switch it on in the accessibility settings 04:25.880 --> 04:27.360 under spoken content. 04:27.360 --> 04:30.120 Then you swipe down from the top of the screen 04:30.120 --> 04:30.900 with two fingers, 04:31.320 --> 04:33.440 and the phone reads out what's in front of you. 04:34.080 --> 04:34.960 And on Android? 04:35.600 --> 04:36.600 Also in accessibility, 04:37.200 --> 04:39.100 only every manufacturer calls it 04:39.100 --> 04:40.160 something slightly different. 04:40.780 --> 04:41.700 You're looking for the feature 04:41.700 --> 04:43.960 that reads selected text aloud. 04:44.820 --> 04:47.540 I've seen something in there called talkback. 04:48.240 --> 04:49.820 And here I have to warn you, 04:50.140 --> 04:51.340 because this is a trap 04:51.340 --> 04:53.260 an awful lot of people fall into. 04:54.120 --> 04:56.400 Talkback isn't reading text out. 04:56.400 --> 04:58.400 It's a full screen reader 04:58.400 --> 05:00.700 for people who can't see the screen at all. 05:01.360 --> 05:01.620 Meaning? 05:02.280 --> 05:03.640 Meaning that once it's on, 05:03.980 --> 05:06.120 the whole way you use the phone changes. 05:06.900 --> 05:08.500 One tap doesn't tap anymore. 05:08.980 --> 05:09.920 It selects. 05:10.440 --> 05:11.000 To tap, 05:11.240 --> 05:12.420 you have to touch twice. 05:12.980 --> 05:15.560 So Gran switches it on in good faith, 05:15.740 --> 05:16.340 and... 05:16.340 --> 05:18.100 And the phone stops doing as it's told, 05:18.480 --> 05:20.180 and she doesn't know how to turn it off. 05:20.620 --> 05:21.660 I've seen it happen. 05:21.660 --> 05:23.660 If you can see the print, 05:23.820 --> 05:25.120 and it's simply too small, 05:25.660 --> 05:27.320 talkback isn't what you're after. 05:28.120 --> 05:29.300 Good thing you've said that, 05:29.520 --> 05:31.440 because I'd have switched it on for her. 05:31.880 --> 05:32.480 I know. 05:32.820 --> 05:33.520 We all would. 05:34.820 --> 05:35.820 And the fourth route. 05:36.300 --> 05:37.240 The pharmacist. 05:37.900 --> 05:40.360 The oldest technology in the world. 05:40.960 --> 05:42.220 And the only one where somebody 05:42.220 --> 05:43.360 will not just read it, 05:43.580 --> 05:45.060 but can explain it. 05:45.480 --> 05:46.640 No app does that. 05:47.100 --> 05:47.980 Mine doesn't either. 05:47.980 --> 05:50.280 But let's not kid ourselves. 05:50.920 --> 05:52.340 It rather depends whether there's 05:52.340 --> 05:54.200 half the street queuing behind you. 05:54.820 --> 05:55.220 It does. 05:55.860 --> 05:57.520 How far that conversation goes 05:57.520 --> 05:58.720 depends on the pharmacy, 05:59.080 --> 05:59.820 the time of day, 06:00.200 --> 06:00.640 the queue, 06:01.080 --> 06:02.680 which is why it's worth turning up 06:02.680 --> 06:03.940 with a specific question 06:03.940 --> 06:05.480 and with a list of the medicines. 06:06.320 --> 06:07.760 Then the conversation is shorter, 06:08.160 --> 06:09.480 and you get more out of it. 06:10.120 --> 06:10.700 Right then. 06:11.300 --> 06:13.640 And where does your app come into all this? 06:13.760 --> 06:14.860 At the end of that queue, 06:15.180 --> 06:15.940 not the start. 06:15.940 --> 06:18.780 And now I'm going to say something 06:18.780 --> 06:21.120 it would be easy to keep quiet about. 06:21.940 --> 06:24.180 And keeping quiet would be dishonest. 06:24.740 --> 06:25.180 Go on. 06:25.720 --> 06:27.080 The Moya Potechka app 06:27.080 --> 06:30.520 does not read the leaflet aloud. 06:31.140 --> 06:31.700 Hang on. 06:32.440 --> 06:34.700 A whole episode about the small print 06:34.700 --> 06:35.520 on a leaflet, 06:35.740 --> 06:37.960 and your app doesn't read it. 06:38.320 --> 06:38.900 It doesn't. 06:39.400 --> 06:40.900 All it does with the leaflet 06:40.900 --> 06:41.760 is open it, 06:42.080 --> 06:43.620 as a link to the official register 06:43.620 --> 06:44.440 in the browser. 06:45.440 --> 06:46.360 What it reads aloud 06:46.360 --> 06:49.160 is the Summary of Product Characteristics, 06:49.920 --> 06:52.000 SMPC for short. 06:52.420 --> 06:54.680 And that's something other than the leaflet? 06:55.080 --> 06:56.120 It's a different document. 06:56.540 --> 06:57.300 An official one, 06:57.560 --> 06:58.580 about the same medicine, 06:58.860 --> 06:59.920 covering the same ground, 07:00.460 --> 07:02.860 but written for doctors and pharmacists, 07:03.260 --> 07:03.900 not for you. 07:03.900 --> 07:06.240 So you swap the document over 07:06.240 --> 07:08.440 and tell me you've solved Gran's problem? 07:08.940 --> 07:10.240 I knew you'd say that. 07:10.900 --> 07:13.900 And that's exactly the charge I put to myself. 07:14.240 --> 07:15.700 And what did you tell yourself? 07:16.100 --> 07:17.780 That I have to say it plainly, 07:18.140 --> 07:19.720 instead of hoping nobody notices. 07:20.680 --> 07:23.040 The Summary is a harder text. 07:23.040 --> 07:26.180 I'm not pretending it's an easier one. 07:26.920 --> 07:27.880 Hearing it read aloud 07:27.880 --> 07:29.520 doesn't make it any simpler. 07:30.220 --> 07:31.940 So why listen to it at all? 07:32.420 --> 07:33.720 Because in the app, 07:33.980 --> 07:34.820 it's text, 07:35.260 --> 07:36.660 not a sheet of paper. 07:37.540 --> 07:39.100 And text you can enlarge, 07:39.420 --> 07:39.740 translate, 07:40.260 --> 07:41.180 and have read aloud. 07:42.120 --> 07:43.020 Paper you can't. 07:43.840 --> 07:44.800 That's the whole advantage. 07:45.560 --> 07:46.180 One thing, 07:46.620 --> 07:47.980 but a fundamental one. 07:48.320 --> 07:48.900 All right. 07:49.480 --> 07:51.360 That's an honest way of putting it. 07:51.360 --> 07:53.100 How it works in practice. 07:54.160 --> 07:56.000 On the medicine's documentation screen, 07:56.360 --> 07:57.480 there's a speaker button. 07:58.480 --> 07:59.040 One tap, 07:59.380 --> 08:01.160 and the phone reads the name of the medicine, 08:01.600 --> 08:02.460 and then the sections, 08:02.820 --> 08:03.580 one after another. 08:04.240 --> 08:05.860 The same button stops it, 08:06.080 --> 08:06.960 and starts it again. 08:07.380 --> 08:08.760 How many sections are there? 08:09.300 --> 08:09.600 Nine. 08:10.280 --> 08:10.880 Indications, 08:11.460 --> 08:11.980 dosage, 08:12.440 --> 08:13.440 contraindications, 08:14.080 --> 08:14.400 warnings, 08:15.200 --> 08:15.760 interactions, 08:16.580 --> 08:18.000 pregnancy and breastfeeding, 08:18.780 --> 08:19.420 side effects, 08:20.000 --> 08:20.300 overdose, 08:20.300 --> 08:21.480 and storage. 08:21.920 --> 08:23.200 That's a lot of listening. 08:23.480 --> 08:24.080 Which is why, 08:24.340 --> 08:25.920 if you expand particular sections, 08:26.320 --> 08:27.600 you'll hear only those. 08:28.260 --> 08:29.920 And if you don't expand any of them, 08:30.320 --> 08:32.480 the app reads out all the sections it has. 08:32.900 --> 08:34.180 What about the internet? 08:34.740 --> 08:38.180 Because the signal at my mum's can be hit and miss. 08:38.180 --> 08:41.520 The voice itself comes from the speech engine in the phone. 08:42.120 --> 08:43.560 Nothing goes off to the network. 08:44.200 --> 08:45.440 The documentation, though, 08:45.580 --> 08:46.760 has to sync once. 08:47.620 --> 08:48.840 After that first sync, 08:49.140 --> 08:50.540 you can listen to it in flight mode, 08:50.740 --> 08:51.220 if you like. 08:51.860 --> 08:52.800 Does it cost anything? 08:53.360 --> 08:53.880 No. 08:54.200 --> 08:55.400 It works on the free plan. 08:56.400 --> 08:58.620 Now tell me what it doesn't do, 08:58.980 --> 09:00.420 because I know you, 09:00.560 --> 09:02.380 and I know you've got a list. 09:02.380 --> 09:04.060 I have. 09:04.560 --> 09:05.920 It doesn't summarise, 09:06.060 --> 09:07.160 and it doesn't simplify. 09:08.100 --> 09:10.380 It reads the official text word for word. 09:11.020 --> 09:12.620 It doesn't put it into plain English, 09:12.880 --> 09:14.060 and it doesn't shorten it, 09:14.440 --> 09:17.440 because shortening is where twisting the content begins. 09:17.900 --> 09:19.680 It doesn't advise. 09:20.060 --> 09:21.440 It doesn't advise. 09:22.000 --> 09:24.640 It won't judge whether this medicine is right for you, 09:25.060 --> 09:26.120 it won't pick a dose, 09:26.380 --> 09:29.540 and it won't tell you whether you can take it alongside something else. 09:29.540 --> 09:32.640 That's what the GP and the pharmacist are for. 09:33.620 --> 09:35.540 And if somebody can't see at all... 09:36.100 --> 09:38.680 And here I'll say something you don't normally say on a podcast 09:38.680 --> 09:40.000 about your own product. 09:40.800 --> 09:43.300 I haven't been through this whole path with a blind person, 09:43.840 --> 09:46.100 from adding a medicine through the documentation 09:46.100 --> 09:47.780 to listening to a section. 09:48.700 --> 09:50.400 I know the button has state descriptions, 09:51.040 --> 09:53.860 and that the speed and the voice come from the phone's system settings. 09:54.380 --> 09:56.380 But I haven't tested the whole path, 09:56.820 --> 09:58.660 and I'm not going to pretend I have. 09:58.980 --> 09:59.460 Right. 10:00.000 --> 10:01.620 That does sound honest. 10:02.580 --> 10:05.220 If somebody uses a screen reader and something falls over, 10:05.640 --> 10:06.580 please get in touch. 10:07.100 --> 10:09.920 It genuinely would change the order I work on things. 10:10.840 --> 10:13.800 There's one more situation that looks just the same, 10:14.060 --> 10:16.020 and has nothing to do with eyesight. 10:16.740 --> 10:17.180 Go on. 10:17.280 --> 10:20.120 The lady who helps with my mum speaks Ukrainian. 10:21.080 --> 10:22.080 Mum doesn't. 10:22.480 --> 10:25.920 And the medicine's documentation isn't in Ukrainian. 10:25.920 --> 10:29.600 A section can be translated, on the device itself, 10:29.600 --> 10:31.300 into one of three languages, 10:31.860 --> 10:34.140 English, Ukrainian or Russian. 10:35.020 --> 10:36.400 And once a section is translated, 10:36.900 --> 10:39.500 the voice switches to the language you can see on the screen. 10:39.500 --> 10:42.020 So she hears what she sees. 10:42.540 --> 10:45.180 That's the rule, and there's no exception to it. 10:45.180 --> 10:48.320 The app reads exactly what's on the screen. 10:48.320 --> 10:49.980 You're looking at the original, 10:50.480 --> 10:51.440 you hear the original. 10:51.900 --> 10:53.200 You're looking at the translation, 10:53.840 --> 10:55.220 you hear the translation. 10:55.220 --> 10:58.500 You'll never hear content you can't see. 10:58.900 --> 11:02.020 You know what the saddest part of all this is? 11:02.940 --> 11:07.240 That for years, my mum simply took them the way she'd always taken them. 11:07.900 --> 11:09.740 Not because she didn't care. 11:09.740 --> 11:13.020 Only because the bit of paper was too small. 11:13.660 --> 11:14.060 Yes. 11:14.840 --> 11:15.980 And that's the nub of it. 11:16.460 --> 11:18.660 The small print isn't going anywhere. 11:19.540 --> 11:22.360 But it isn't the only form this content exists in. 11:23.000 --> 11:23.720 There's the leaflet. 11:24.060 --> 11:24.840 There's the file. 11:25.400 --> 11:26.600 There's the official summary. 11:27.280 --> 11:29.380 And there's a person behind the counter. 11:30.180 --> 11:33.040 You only need one of those routes to be open to you. 11:33.040 --> 11:35.680 If you'd like to give it a go, 11:36.200 --> 11:37.620 that's the Moya Potechka app. 11:38.180 --> 11:40.060 The app is free and works offline. 11:40.840 --> 11:42.540 You'll find the link in the episode description. 11:43.620 --> 11:44.920 And if you enjoyed the episode, 11:45.460 --> 11:48.020 share it with someone who's sitting in the kitchen under a lamp, 11:48.340 --> 11:49.880 fighting with a magnifying glass. 11:50.800 --> 11:51.980 It is funny, though. 11:52.620 --> 11:55.000 A whole episode about reading aloud, 11:55.480 --> 11:57.800 told by two voices that... 11:57.800 --> 12:00.280 That's exactly where I was going. 12:00.280 --> 12:01.880 One more thing before we go. 12:01.880 --> 12:04.320 This podcast is made with the help of artificial intelligence. 12:04.320 --> 12:09.740 I wrote the dialogue myself, but the voices you're hearing aren't ours. 12:09.740 --> 12:16.800 They're AI-generated synthetic voices. We're telling you this straight out, 12:16.800 --> 12:23.180 because we think it should be the standard for any podcast made this way. 12:23.180 --> 12:25.300 Full details are in the show notes and in every episode description. 12:25.300 --> 12:26.600 Until next time. 12:26.600 --> 12:29.100 Until next time.