England and Wales Court of Appeal (Criminal Division) Decisions

IN THE COURT OF APPEAL (CRIMINAL DIVISION)
ON APPEAL FROM CROWN COURT AT WOOLWICH
Her Honour Judge Robinson
T20111073

		Royal Courts of Justice Strand, London, WC2A 2LL
		18/11/2015

Lady Justice Rafferty:

1. On 12 October 2012 in the Crown Court at Woolwich the appellant Christopher Brown was convicted of 6 counts (2-3 and 9-12) of making indecent images of children contrary to s 1 (1) (a) Protection of Children Act 1978. On 15 November 2012 he was sentenced to a suspended sentence order on each count concurrent. He was acquitted of 6 further counts of making indecent images of children (Counts 1 and 4-8) and of 2 counts of attempting to make indecent images of children (13-14). He appeals with the leave of the Full Court.
2. On 20 April 2012 the appellant was arrested as a result of police monitoring of a peer-to-peer file sharing computer application "Frostwire." Indecent images of children were on his computer. In interview he admitted installing Frostwire. He said he had inadvertently accessed indecent images of children whilst trying to download mainstream pornography then deleted or tried to delete the images of children. He sometimes left the computer on all night for a download and the indecent images had come through accidentally. He made no comment in a subsequent interview.
3. The Crown's case was that he had viewed all the indecent images. It relied on the expert Simon Field who concluded that some images had been deleted or were not accessible to the ordinary user. The appellant was acquitted of these counts. However, an automatically generated folder, "thumbcache", was invisible and inaccessible and the ordinary user of a computer would not know of its existence. The appellant was convicted on these counts as well as a rolled-up count (12) relating to the balance of the images.
4. The Crown contended that the file titles made it obvious they contained indecent images of children and that some had clearly been selected for download as a result of his searches. Searches by Mr Field using terms supplied by the appellant did not result in indecent images of children. The Crown thus suggested that the appellant had actively sought out indecent images of children.
5. The defence was that the downloads were accidental and he had deleted all he could. He was of previous good character. He had reconfigured the software so it could not automatically connect to the internet and incomplete files could no longer be automatically downloaded. Mr Field conceded that the filenames were not such that from them the contents could be discerned and one could not tell what would appear on the screen before the images were opened.
6. The appellant in interview said he worked for solicitors in their IT department. At a further interview after more detailed examination of the hard drive he read a prepared statement in which he said that he had fully commented in his first interview and had nothing to add.
7. Stephen Faulkner, police forensic computer analyst, found associated with "Frostwire" 10 indecent images with filenames indicative of child abuse. 10 images of adult pornography in the same folder were accessible to someone looking at the contents of "Frostwire." Many files contained names indicative of child abuse material but it was not possible to view them because they had not fully downloaded. There was evidence files had been accessed, and evidence of their names, but he could not find them. It was possible they had been moved, deleted or renamed. He found about 500 adult pornographic images and a further 8 indecent images of children.

Expert evidence

8. The overwhelming inference was that all images came from the use of Frostwire whose basic operation was agreed. It enabled "person-to-person" sharing, as opposed to accessing a file from a central server. An user opened Frostwire, typed in a search term and clicked "search." He would see a list of files and filenames on other computers presently connected to the internet and whose names matched his search terms. He would click "download" on files he wanted and downloading would begin, if in large number then in tranches. As soon as "download" was clicked the file would go into a folder marked "incomplete" and once fully downloaded would leave it and be saved in "new folder" created in response to prompts. Once the download began the user would have the option of more searches.
9. After the appellant installed Frostwire on 28th March and created the new folder, the downloads were saved in it. The default configuration meant Frostwire connected to the internet as soon as the computer was switched on. He altered that default setting so that it connected to the internet only were Frostwire opened though the experts could not say when he did so.
10. For the defence Dr Tipe's undisputed evidence was that this was consistent with his trying to prevent further downloading. On 28/29th March of four files viewed in one session, one from "incomplete" three from "new", three had names indicative of indecent images of children, the fourth indicated both child and adult pornography. All had been deleted.
11. The virtual store ("VS") was created by the computer when an user first installed Frostwire. Were the user to delete "incomplete" and "new" subfolders and reopen Frostwire, it tried to create a new folder to store downloads in a location protected by its operating system so as to prevent accidental deletion. This new folder was created in the VS, hidden from the user unless actively sought.
12. Mr Field said one could access files in the VS without specialist software. Dr Tipe said the appellant could not have looked using Windows Explorer without first altering the computer settings, viewing, deleting, then changing the settings back to the default settings. This was consistent with his account in interview that he clicked on Explorer to view a folder he believed to be in Frostwire and was prompted that the file did not exist.
13. There was no dispute that if one opened Frostwire one could access all files, as the appellant had done several times, including on 31st March when in the VS he viewed three files whose names suggested indecent images of children. Dr Tipe said that whilst it was not possible to access those files via Windows Explorer they could be accessed by running Frostwire and double clicking on the filename.
14. On 17th April Frostwire was started at 2250 and images downloaded (Counts 1 and 7-8, of which the appellant was acquitted). They included two of children, opened or viewed and not deleted, pleaded in Count 12. Dr Tipe said they had been viewed in the same way as the three on 31st March, by double clicking on the file in Frostwire.
15. The two files viewed on 17th April but not deleted, on the face of it not consistent with his account that he had deleted all images of children, prompted further tests. (There was no evidence he had tried to delete the files and the experts agreed an user did not have to open a file to delete it). Frostwire automatically made any downloaded file available for sharing assuming the other person was connected to the internet and the search term was "Return one of your files". It was not possible to delete a file in the process of being shared.
16. On 18th April his account was connected to the internet but did not start Frostwire. On 19th April it did and files were selected for download which was not completed. Mr Field felt ninety-nine files names suggested images of children but others referred to adult or gay pornography. The incomplete download could have been for a number of reasons - loss of internet connectivity, the user stopping the download or closing Frostwire.
17. The effect of the expert evidence was that the appellant selected "all files" for download on 17th and 19th April. His account was that he had not clicked "download" so he assumed a malfunction but neither expert was invited to deal with this. Mr Field said twenty images in the VS, fully downloaded on 17th April but not deleted, had been selected for download at 23.03. Fewer than seventeen were indecent images of children. Ten were pleaded in Count 12. There was no evidence any had been viewed.
18. In Thumbcache, a location upon he computer to which deleted gallery images were sent, his account held more than eleven hundred images ranging in dimension from postage stamp to 2" square. Of the one hundred and sixty with sexual content seventy-five were of children.
19. The images were in a named folder which, when opened, automatically put some images on the screen in thumbnail form. Until a folder was opened one did not know which ones would be displayed on the screen. If it held a large number only a few would be displayed. Were a folder deleted the only images entering thumbcache were those which had been displayed. One could delete a file containing an image without opening it.

Areas of dispute

20. There were two areas of dispute between the experts: which search terms might have been used by the appellant and what conclusions could be drawn from the file names. Dr Tipe's view was that his searches were consistent with the appellant's account and that it was not possible to be sure that he searched exclusively pre-teen material. Mr Field, who had not precisely replicated the search as done by the appellant, was less confident.
21. Mr Field agreed that some file names were designed to mislead but if a filename suggested an indecent image of a child he would not risk downloading it. Dr Tipe told the jury that files shared person-to-person often had sexually explicit names but unrelated content. Distributors employed a scattergun approach in the hope of more downloads. The names of some of the files in the instant case contained references to underage, overage and gay sexual content, so that one did not know what would be in particular files.

Ruling on submission at the close of the case for the Crown

22. The judge reminded herself of R v Smith and Jayson [2002] EWCA Crim 683 on what constituted a making for the purpose of s 1 (1) (a) Protection of Children Act 1978. On Counts 1-12 the downloaded images fell into three categories described as accessible, inaccessible and a thumbcache. The files in the VS were created and downloaded intentionally. The issue was whether he downloaded them with the requisite knowledge. On all the evidence, including a lot of file names indicative of indecent images of a child and two files viewed containing such, a jury could be sure [of guilt]. That he was seeking to download adult pornography and could not be sure of the content of the file without opening it was not sufficient for her to withdraw the case. Material was described as inaccessible because images downloaded but deleted made it impossible to say when they had been downloaded, by whom, the file name or where it came from. There was therefore evidence that the image had been intentionally downloaded. The jury would be entitled to infer he was aware of the likely content because they must have come from Frostwire and were likely to be named similarly to the others. It was common ground that the thumbcache images were sufficiently clear on the screen to enable one to see the image. True, there was no evidence he was aware of the thumbcache but there was evidence on which the jury could be sure of a deliberate downloading and that the files contained indecent images of children.

The defence case

23. The appellant, 30 and of good character, married with children, told the jury his IT knowledge was that of the average home user. In March/April 2011 he installed "Frostwire" to download and view adult pornography without internet connection. On 28^th March he searched using terms such as "teen," "fuck" and "girlfriend" which would give results for adult pornography. He selected "all" results, clicked "download" and left the files to download whilst he did another search. He was not present during all downloads. He followed the Frostwire "wizard" and created a new folder. When he noticed the results of the initial search he thought he had done something wrong and deleted whole and incomplete folders.
24. He had opened four files but notwithstanding their names none held indecent images. He deleted everything as a result of what he had seen. On 31st March he tried Frostwire again but when he looked at the pictures realised it had not downloaded what he had expected. He stopped using his computer. All thumbnails in thumbcache had been deleted between 28th and 31st March. He had viewed the images very briefly once he realised what they were and did not study the file names. Duplicate thumbcache images were not from anything he had done. He opened three files on 31^st March via Frostwire, which took one to Windows Explorer, and displayed in thumbnail what had been downloaded, exactly as on 28-29th March. He could not explain how he had viewed the first four files in the original new folder created by Frostwire whereas the three viewed on 31st March were in the VS which could not be accessed via Windows Explorer. He assumed it was to do with the computer. He had not used it again until 17th April.
25. His wife on 17th April found it slow and temperamental. He first realised it was set to download files automatically when he saw a tiny icon showing Frostwire was "trying to do stuff." and he disabled auto-connect. Commenting on the experts' agreement that the files created on 17th April must have been selected to be downloaded he said "Computers don't always do what you want, otherwise you wouldn't have experts." He did not know why Frostwire had not begun downloading earlier in the evening when his wife had briefly connected to the internet, but files had not begun to download until two minutes after the internet connection started.
26. He did not know of the VS or of thumbcache. He double-clicked on the files via Internet Explorer but could not find them so as to delete them. He accepted that in the process two images of children came up. He had been trying to delete all images when interrupted. He had not intentionally searched for, downloaded or made indecent images of children.
27. On 19th April he opened Frostwire to delete any remaining images but failed. In interview he told the truth. In the second interview he followed advice.

Grounds of appeal

28. Grounds of appeal advanced by Mr James Wood QC who did not appear below are that the guilty verdicts were logically inconsistent with the acquittals. The judge erred in written directions on Counts 2-3 and 9-12, wrongly asserting that the folder containing those images must have been opened because they were viewed on screen and therefore downloaded and viewed intentionally rather than by accident and the issue was whether he knew the files did or were likely to contain an indecent image of a child, whereas the true issue was whether he had deliberately downloaded and viewed the images. The direction that the issue was "…the circumstances in which[the images] came to be on the computer in the first place…" was a serious misdirection; the appellant was charged with "making" the thumbcache images and the Crown thus had to prove that he knew the images were automatically saved. The essential elements of the offences as particularised had to be made out in connection with the particular "making" alleged. The indicted "making" was an unknowing and unintentional act and the jury was never properly directed as to it.
29. Whether the images had been made and viewed intentionally should have been left to the jury. An image was not "made" in a deliberate and intentional act but by an unknown automatic process of the computer which generated the image in thumbcache. The second limb of the direction did not address the timing of when the appellant knew the files did or were likely to contain an indecent image of a child. The knowledge had to be at the time of the "making." On the evidence this was not upon entry into Frostwire of search terms generating automated download but when the file names were listed and thumbnails simultaneously and automatically stored in thumbcache.

Grounds of opposition

30. For the respondent Crown Mr Moll who also appeared below relied on what he described as a simple factual analysis: The appellant deliberately searched for indecent images of children and knew the images had such a content when he opened them. There was a clear consensus that he had viewed then deleted the thumbcache images and the verdicts were plainly centred upon where on the computer the particular indecent images of children were.
31. He reminded us of the appellant's evidence "Once I had viewed and seen what was there it was only briefly on the screen…I went "Oh my God" and got rid of it." The agreed evidence was that the thumbcaches had been viewed at some stage whereas there was only an inference the others had been viewed
32. The images in Counts 1, 7 and 8 and the attempts in Counts 13-14 had been downloaded or selected for download on 17th April when, he said, his computer was malfunctioning and was downloading automatically. The experts disagreed as to whether files titled to suggest indecent images of children would in fact contain such (Counts 13-14). There was no evidence that the items in the unallocated space (Counts 4-6) had been viewed or even how they came to be on the computer in the first place. There was a logical basis for the verdicts.
33. The judge directed the jury on the Counts of which the appellant was acquitted that it had to be sure he intentionally downloaded the files and knew they contained or were likely to contain an indecent image of a child. There were seventy-five indecent images of children in the thumbcache and having viewed a few he must have known they contained such images. The issue was not whether he deliberately downloaded and viewed the images. His evidence was that he had. There was ample evidence from which jury could infer he knew exactly what he was downloading.

Discussion and conclusion

34. A good deal of attention was devoted in written submissions, for which we were grateful, to technical aspects of how the computer and any relevant software worked. Oral advocacy revealed that the issue was far narrower and did not require detailed analysis of technology. The first and most powerful of Mr Wood's arguments, that the verdicts defied logic, proved dispositive and we have not found it necessary to reach a conclusion on the balance of the submissions.
35. Although the indictment was not specific as to the dates when material was downloaded some events could be dated. as we have already set out.
36. On 31^st March Frostwire settings were changed so that files were stored in the VS, created by the computer itself in the systems folders which, on the evidence of Dr Tipe, ordinary users cannot access for fear of wrecking the workings of the computer. Mr Field disagreed,. Whichever expert view did prevail or should have prevailed, three files were viewed and deleted.
37. The VS counts (1, 7 and 8) were left to the jury on the basis that the files were downloaded on 17^th April, not deleted, and found in the VS. Summing up the judge said:

"The issues which you have to decide in respect of these three counts of these. Are you sure the defendant downloaded these files intentionally?… Are you sure that the defendant knew the files did or were likely to contain an indecent image of a child? If the answer to both these questions is yes the defendant is guilty. If the answer to either question is no, he's not guilty."

The jury returned verdicts of not guilty.

38. The thumbcache counts (2, 3, 9, 10 and 11) were left on the basis that the files had been downloaded viewed and deleted. The judge said:

"There is no dispute that these were downloaded by the defendant using FrostWire. Further, the folder they were contained in must have been opened because they were viewed on screen. Therefore, they were downloaded and viewed intentionally rather than by accident. Therefore in respect of these counts there is only one issue for you to decide which is this: are you sure the defendant knew the files did or were likely to contain an indecent image of a child? If the answer that question is yes, the defendant is guilty if the answer is no, he is not guilty."

The jury returned verdicts of guilty.

39. As to the unallocated space counts (4, 5 and 6) the judge said:

"Although there is no evidence as to the original filename for these images, the date or time and there were downloaded or by which user account, the defendant accepts that they can only have been downloaded by him via FrostWire. Because there is no evidence as to how these files came to be downloaded the prosecution seeks to draw an inference from all of the circumstances that they were downloaded intentionally and with knowledge as to the likely contents………….So, again, the issues in respect of these three counts which you have to decide are as follows. First are you sure that the defendant downloaded these files intentionally and, secondly, are you sure that the defendant knew they did all were likely to contain an indecent image of a child?"

The jury returned verdicts of not guilty.

40. As to the attempt counts (13 and 14) the judge said:

"The issues which you have to decide in respect of these two counts are as follows. Are you sure that the defendant intended to download an image which she knew did contain all was likely to contain an indecent image of a child and, secondly, are you sure………….

The jury returned verdicts of not guilty.

41. The rolled-up count (12) encapsulated the balance of the downloaded images in whatever category. The judge said:

"So the issues arise on this count as in relation to counts 1 to 11 for each of those three locations. If you are sure that one or more of the images, the subject of this count, was downloaded or viewed on the screen by the defendant intentionally and with the knowledge that the file did or was likely to contain an indecent image of a child then the defendant is guilty on count 12. It is not necessary for the prosecution to prove its case as to all 131 images referred to in the indictment."

The jury returned verdicts of guilty, which could have reflected only the thumbcache images within that catch-all category.

42. Mr Wood QC argued, as to the thumbcache, that the making was the result of the downloading, not of the viewing. It was the state of mind of the appellant at the point of downloading not at the point of viewing which was relevant, whether in thumbnail or full screen format.
43. The primary difficulty for the respondent Crown lies in identifying a rational reconciliation between the verdicts of not guilty and of guilty. They are not explicable by reference to the time at which the thumbcache images were downloaded. The Crown argued (contrary to the case it led at trial) that a point must have been reached by which the appellant must have known that the files were likely to contain indecent images. Leaving aside how matters were led below, the difficulties for the Crown are the acquittals on counts 13 and 14 as well as those on counts 1, 7 and 8, which related to attempted and actual downloads in April. In any event, as it conceded, it is impossible to give a date when the thumbcache images, or any of them, were downloaded: all thumbcache images might have been downloaded in the first download in the period 28-31 March and none later
44. Nor can the verdicts be explained by the appellant having viewed these images. He had also viewed images other than those in the thumbcache but been acquitted as to them. The proposition, not advanced below, that the thumbcache images may have been the same as the "other" images, was not supported on the evidence. The thumbcache images were clearly left to the jury on the basis that they were distinct from the "other" images and it was not open to the jury to convict on any other basis.
45. There was clear evidence that when the appellant saw indecent material, he deleted it. This was as true in respect of the "other" images as of the thumbcache images. Indeed the evidence was plain that these images were only created in this format as a consequence of deletion.
46. In the course of submissions counsel for the respondent Crown was invited to offer a reasoned explanation for these apparently inconsistent verdicts. He did his valiant best but was unable to help us. We had little difficulty identifying irrationality and inconsistency but had no success when we sought to understand any possible logic in the verdicts.
47. Thus, as we indicated, these apparently complex arguments, although stratified and requiring close attention to detail, revealed a ground of appeal familiar and simply stated. It was unnecessary for us to attempt a masterclass in computer software analysis or revisit the meaning of statutory terms. It was enough to conclude, on these facts, that the verdicts were inconsistent.
48. This appeal is allowed.